Chapter 1

Introduction to Operations Management

1.1 Introduction
Manufacturing or service operation is an integrated extension of corporate policy to the
extent that the total enterprise exists and is organized as a system of transformation (value
adding) activities to satisfy pluralistic needs and expectations of stake holders. Production
is the process of transforming a variety of resources into goods or services.
Transformation is not merely conversion of inputs into outputs but also, significant value
addition to inputs, by the processes, to create utilities to the end users. esources include
materials, machines, human time, and energy and so on. The output of the production
processes may be manufactured goods, such as automobiles and computers, or services
such as health care or financial transactions. The term !operations" describes the #hole set
of activities associated #ith the production of services and goods. The system approach is
illustrated in $igure %.%.
Materials
&abour 'oods and (ervices

)*uipment
+apital
$)),-.+/ 01$2M.T021 3 +21T2&
$igure %.%. (ystem diagram for operations management
The operations management is about managing the conversion process. 0n vie# of the
global competition, it is essential to provide goods and services at competitive prices. 0n
production4operations management, one can see a lot of scope to minimize various costs,
since a significant portion of the product cost is incurred in this stage of the business. The
production systems are classified based on nature of products and type of flo#s. 0t is best
illustrated in Table %.%.
Table %.%.+lassification of production system
BASIS CLASSIFICATION EXAMLES
Type of output Products +onsumer goods like
furniture, T5, radio and
producer goods like lathe,
%
T.1($2M.T021 P2+)((
01P6T 26TP6T
milling machines
Type of output (ervices 7ealth, hospitality, airlines,
banking education system
Type of flo# Pro8ects +onstruction of dams,
bridges,
Type of flo# 9ob shop Machine shop, auto repair,
hospital
Type of flo# $lo# shop or mass
production
+ar assembly, T5, shoe
factory
Type of flo# +ontinuous process Postal services, chemical
plant, thermal po#er plant
1.!. "ecision areas in operations management#
:ithin the operations function, management decisions can be divided into three broad
areas.
%. (trategic (long;term) decisions
<. Tactical (intermediate;term) decisions
=. 2perational planning and control (short;term) decisions
Table %.<. ,ecision areas in operations management
Strategic Tactica$ Operationa$
Product design &abors re*uired ,ay to day scheduling
&ocation of facilities .ggregate planning .ssignment of 9obs to
machine
+apacity decisions )xtra shift. 2ver time (e*uencing of 9obs
+apacity expansion ,elivery planning (ome inventory decisions
&ayout decisions 0nventory carrying
Process planning .ggregate planning

1.!.1 Current issues in operations management>
%. )ffectively consolidating the operations resulting from mergers.
<. ,eveloping flexible supply chains to enable mass customization of products and
services.
=. Managing global supplier, production and distribution net#orks
?. 0ncreased commoditization of suppliers
@. .chieving the service factory
A. .chieving good service from service firms
1.!.! %h& Operations Management'
%. . business education is incomplete #ithout an understanding of modern
approaches to managing operations.
<. 2perations management provides a systematic #ay of looking at organizational
processes.
<
=. The concepts and tools of 2M are #idely used in managing other functions of a
business.
?. 2perations Management presents interesting career opportunities.
1.(. Manu)acturing and ser*ice operation
Manufacturing implies production of a tangible output, such as automobiles, a clock, a
radio, a golf ball, a refrigerator or something similar that #e can see or touch. (ervice on
the other hand, generally implies an act rather than a physical product.
These t#o are similar in terms of #hat is done but different in terms of ho# it is done.
-oth involve design and operating decisions
Manufacturer must decide ho# large a factory is needed and service
organizations must decide ho# large a building they need (7ospital)
-oth must take location decisions and both become involved in scheduling and
controlling operations and allocating scarce resources.
-ut the differences involve the follo#ing>
The nature and consumption of output
6niformity of output
&abour content of 8ob uniformity of output
Measurement of productivity
Table %.=. ,ifferences bet#een production and service operations
roduct+operations Ser*ice+operations
Tangible 0ntangible
Produce, store and sell Production and consumption simultaneous
mostly
1o customer involvement during
production process customer is not part of
the process)
+ustomer presence and involvement is
essential (customer is part of the process)
More automation (less labour intensive) &ess automation (more labour intensive)
Buality control is easy Buality control is very difficult and *uality
is sub8ective
+an be stored for future demand +an not be stored
Productivity;ob8ective Productivity;sub8ective
+ustomization;difficult +ustomization is easy
6tility;mostly physical Time, place, form, or psychological utility
Products can be returned if you are not
satisfied
(ervices are consumed and can not be
returned
1.(.1 "esign o) Ser*ices#
:hen you design a service product, the follo#ing points should be considered>
=
%. To #hat extend #ill the customer be involved in the processC
<. 7o# *uickly #ill service be providedC
=. 7o# standardized or customized #ill the service beC
?. :hat variety of services #ill be offeredC
@. :hat geographical area #ill be servedC
1.(.! "esign o) roduct#
Product design should consider the follo#ing points>
%. 0t has direct bearing on plant layout and in;process material flo#
<. .lternative plans to produce (processes)
=. +omponents and substitutes availability
?. )rgonomic considerations
@. ,esign for repairability and redesign
A. ,esign for remanufacturing
D. Modular design
E. ,esign for selling
F. ,esign for making
%G. ,esign for function
Chapter !
Location decisions and mathematica$ mode$s
?
!.1 Introduction
-usiness systems utilize facilities like plant and #arehouses #hile performing the task of
production and distribution of products and services. These decisions are strategic in
nature and it has long;term implication for businesses.
$actors influencing Plant &ocation>
'eneral $actors>
%. .vailability of land
<. .vailability of inputs such as &abour, ra# materials etc.,
=. Proximity to market place
?. .vailability of infrastructure like oads, Ports and rail
@. 'overnment support like 0ncentives, tax holidays etc.,
A. 0ndustrial relations
D. .vailability of community facilities
!.! Mode$s )or Faci$it& $ocation ana$&sis
!.!.1 "imensiona$ Ana$&sis#
0f all the costs #ere tangible and *uantifiable in rupee value, the comparison and
selection of a site is very easy. The least cost site is selected using cost analysis and
break;even analysis. -ut in most cases, there are many intangible costs, #hich may be
better expressed, in relative terms than in absolute terms. $or instance, educational
facility for children or lack of it at a site is difficult to *uantify in absolute terms. -ut
their relative merits can be compared more easily. -ridgeman"s ,imensional analysis
consists in computing these relative merits for each of the cost items, giving each of the
ratios an appropriate #eightage by means of the po#er to #hich it is raised, and
multiplying these #eighted ratios in order to come up #ith a comprehensive figure on the
relative merit of the t#o alternative sites. This method is used to compare t#o sites at a
time. 0f you #ant to compare more sites, one has to go by order. That is by comparing
first t#o sites, then compare the best site #ith next site and so on.
+%
M
, +<
M
, +=
M
HH+z
M
are the different costs associated #ith a site M on the !z"different
cost items.
+%
1
, +<
1
, +=
1
HH+ z
1
are the different costs associated #ith a site 1
:%, :<, :=,H:z are the #eightages given to these cost items,
The relative merit of site M and 1 is given by>
I+%
M
4+%
1
J
:%
K I+<
M
4+<
1
J
:<
K HHI +z
M
4+ z
1
J
:z

0f this is less than 21), it means site 1 is superiorL and vice versa.
'iving appropriate rating using %;%G scale one can compare non;cost factors like
community facilities.
!.!.! Bro,n+-i.son Mode$ )or Site $ocation#
@
This model is more elaborate and considers three classes of site location factors>
%. +ritical; e.g., #ater for chemical processing plant
<. 2b8ective; e.g., transportation costs, distance from existing facilities
=. (ub8ective; e.g., union relationship, community facilities
$or each site !i" a location measure, &Mi M +$Mi K I, K 2$Mi N (%;,) K ($MiJ
:here +$Mi M critical factor meaure for site !i"(it can be either zero or one)
2$Mi M ob8ective factor measure for site !i"(G 2$Mi % and 2$Mi M %)
($Mi M sub8ective factor measure for site !i" (G ($Mi % and ($Mi M %)
, M ob8ective factor decision #eight ( G , %)
(ites #ith higher location measures are preferred to sites #ith lo#er location measures.
!.!.( Center o) -ra*it& Method#
This method is a techni*ue for locating single facilities that considers the existing
facilities. 0t is often used to locate intermediate or distribution #arehouses. This method
begins by placing the existing locations on a coordinate grid system. The purpose is to
establish relative distances bet#een locations. 6sing longitude and latitude coordinates
might be helpful in international decisions. The center of gravity or ne# facility
coordinates is found by calculating K and O coordinates that result in the minimal
transportation cost.
C/ 0 1 2di/3i 45 23i
C& 0 1 2di&3i45 23i
:here
+x M K coordinate of ne# location
+y M O coordinate of ne# location
dix M K coordinate of the 6i
th
7 location
diy M O coordinate of the !i
th
" location
5i M 5olume of goods moved to or from the !i
th
" location
Chapter (
$ant $a&out
=.% 0ntroduction
A
&ayout deals #ith the arrangement of #ork areas and e*uipments inside the plant or
service facility. The basic theme behind the arrangement of #ork areas is to produce the
product economically, to provide the service effectively, and to provide a safe and good
physical environment for the users that is, the #orkers and customers.
(.1.18esu$ts o) -ood p$ant $a&out#
%. )ase of #orking, maximum safety and minimum health hazards for people
<. Minimum handling of materials
=. Minimum damage and spoilage of materials
?. educed congestion of materials, machinery and men
@. $lexibility #ith regard to changing production conditions
(.1.! Basic rincip$es o) La&out#
%. The total movement of materials should be minimum
<. The arrangement of the #ork area should have as much congruence as possible
#ith the flo# pf materials #ithin the plant
=. The layout should ensure ade*uate safety and healthy #orking conditions for the
employees
?. . good layout should take into consideration all the three dimensions of space
available.
@. The layout should be adaptable or flexible enough so as to allo# for probable
changes in the future.
(.1.( Costs associated ,ith p$ant $a&out decisions#
%. +ost of movement of materials
<. +ost of space
=. +ost of production delays
?. +ost of spoilage of materials
@. +ost of customer dissatisfaction due to poor service *uality
A. +osts of labor dissatisfaction and health risks
(.! T&pes o) La&out
%. Product &ayout
<. Process &ayout
=. +ombinational &ayout
?. $ixed Position &ayout
@. +ellular layout
(.!.1roduct $a&out
The e*uipment here is laid out according to the se*uence in #hich it is used for making
the product. 0t is suitable for assembly operations. (.utomobile, +omputer, Tractors).
.dapting a line or product layout makes sense #hen the batch size of a given product or
D
part is large relative to the number of different products or parts produced. .ssembly line
balancing is the techni*ue used design and balance the assembly line layout in most
industry. This layout is laid according to the se*uence of operation. 0t involves lo# set up
time
0t facilitates high
$igure =.%. Product layout
(.!.! rocess $a&out
The most common approach to developing a process layout is to arrange departments
consisting of like processes in a #ay that optimizes their relative placement. This is
typical of the 8ob;shop type of production #here the e*uipment performing similar
operations is grouped together.
)xamples> Machine shop P Milling, &athe, 'rinding, etc.,
roduct B
roduct A
$igure =.<. Process layout
This layout is flexible and can handle variety of 8obs. Material handling is very high in
this layout and #ork in process also high.
E/amp$es# Machine shop, $urniture, 'eneral 7ospitals, 6niversity
Process layout can be designed #ith the ob8ective of reducing movement of materials
bet#een departments. There are computerized layout techni*ues also available like
+.$T, .,&)P etc., .nother method for layout design #idely used in retail and service
firms is systematic layout planning.

(.!.( Com.inationa$ $a&out
This layout tries to combine product and process layout. 0n fact +ellular layout is one
type of 7ybrid or combinational layout.
E
Process 0 Process 00 Process 000 Process 05
(tores (haping ,espatch
&athe operations Painting
Milling
+utting :elding
(.!.9 Fi/ed position $a&out
0t is characterized by a relatively lo# number of production units in comparison #ith
other type of layout. 7ere the e*uipments and people move in and around production
point and product that is manufactured is mostly fixed at one place because of its size.
.ircraft building, (hip building and ocket building are some of the examples.
0
1
,
0
.
$igure =.=. $ixed position layout
(.!.: Ce$$u$ar $a&out or group techno$og&
'roup technology or cellular layout allocates dissimilar machines into cells to #ork on
products that have similar shapes and processing re*uirements. This method is no#
#idely used in metal fabrication, computer chip manufacture, and assembly #ork. The
overall ob8ective is to gain the benefits of product layout in 8ob shop kinds of production.
The benefits are>
%. -etter human relations; +ells consists of a fe# #orkers #ho form a team
<. 0mproved operator expertise;:orkers see only a limited number of different parts
in a finite production cycle, so repetition means *uick learning
=. &ess in;process inventory and material handling
?. $aster production set;up
(teps in developing a +ellular &ayout>
%. 'rouping parts into families that follo# a common se*uence of steps. This step
re*uires developing and maintaining a computerized parts classification and
coding system.
<. 0dentifying dominant flo# pattern of parts families as a basis for location or
relocation of processes.
=. Physically grouping machines and processes into cells.
F
. - +
$igure =.?. +ellular layout
Chapter 9
Linear programming and its app$ication
?.% 0ntroduction
&inear Programming refers to several related mathematical techni*ues that are used to
allocate limited resources among competing demands in an optimal #ay.
&inear programming is an 2perations esearch techni*ue, #hich originated during the
early %F@Gs. 7aving diverse practical applications, this techni*ue has benefited
immensely various organizations in their production and other operations. Prof. '.-.
,antzig is one of the pioneers in formulating the procedure of linear programming.
This techni*ue can be applied in various situations> long range planning, production
planning, #arehousing decisions, supply chain net#orking, distribution decisions,
marketing mix and promotion mix decisions, purchasing decisions and many more
industrial engineering applications.
The basic problem solved by linear programming is that of optimizing either profit or
total costs or some other utility function. 0t takes into consideration the limitations or
constraints on the availability or usage of different resources such as manpo#er,
machinery, materials, time and money as also other limitations and constraints such as
those existing in the market.(only so many units of a product can be sold). &inear
Programming deals #ith optimizing a desired ob8ective under a situation #here there are
various constraints. Most of the management problems are optimal decision;making
problems made under the various limitations.
0llustration>
(uppose a +ompany produces t#o products K and O, both of #hich re*uire a particular
ra# material and a particular machine. Product K re*uires ? machine hours and = kg of
%G
&
M
'
M
:
'
&
'
M
the ra# material per unit of the product and product O re*uires < machine hours and A kg
of the ra# material per unit of the product. (uppose the availability of ra# material and
machine hours is limited. The ra# material is available to the maximum extent of only
<?G kg per month and the machine hours are available to a maximum extent of only <GG
machine hours per month. )ach of the products K and O contribute to the profit margin
by s. D and s. F respectively per unit of the product. 7o# many units of K and O
should this company produce every month to maximize his profitC
$ormulation>
0n the above problem, the company has to decide on the *uantities of K and O #hich are
the decision *aria.$es and this is to .e done so as to ma/imi;e the pro)it margin
,hich is the o.<ecti*e under the $imitations o) resources. &inear programming
problems typically have three elements>
%. ,ecision variables, the determination of #hose value is the problem to be solved
<. 2b8ective function #hich is to be either maximized or minimized
=. +onstraints or limitations related to the decision variables.
The solution to a &PP involves,
%. )xpressing the ob8ective function in an algebraic form involving the decision
variables in algebraic notations. This expression is called the ob8ective function,
#hich is either to be maximized or minimized.
<. )xpressing the constraints in algebraic ine*ualities involving the decision
variables in algebraic notations.
=. The above t#o steps complete the formulation of the linear programming
problem. This can be no# solved for the determination of the optimal values of
the decision variables by means of a mathematical procedure. (impler &PPs can
be solved as a graphical procedure.
The typical operations management applications of linear programming are given belo#.
%. .ggregate production planning> $inding the minimum cost production schedule,
including production rate change costs, given constraints on size of #orkforce and
inventory costs.
<. (ervice productivity analysis> +omparing ho# efficiently different service outlets
are using their resources compared to the best performing unit.
=. Product planning> $inding the optimal product mix #here several products have
different costs and resource re*uirements.
?. Product routing> $inding the optimal routing for a product that must be processed
se*uentially through several machine centers, #ith each machine in a center
having its o#n cost and output characteristics.
@. Process control> Minimizing the amount of scrap material generated by cutting
steel, leather or fabric from a roll or sheet of stock material.
A. 0nventory control> $inding the optimal combination of products to stock in a
#arehouse.
D. ,istribution scheduling> $inding the optimal shipping schedule for distributing
products bet#een factories and #arehouses or retailers
E. (upply chain net#ork design in terms of location of facilities.
%%
There are five essential conditions in a problem situation for linear programming to
pertain. $irst there must be limited resourcesL other#ise there #ould be no problem.
(econd, there must be an explicit ob8ective (such as maximize profit or minimize cost).
Third, there must be linearity. $ourth, there must be homogeneity (the products produced
on a machine are identical, or all the hours available from a #orker are e*ually
productive). $ifth is divisibility.
1ormal linear programming assumes that products and resources can be subdivided into
fractions. 0f this subdivision is not possible, a modification of linear programming called
integer programming can be used.
9.! Linear programming mode$
(ated formally, the linear programming problem entails an optimizing process in #hich
nonnegative values for a set of decision variables K%, K<, K=HHHKn are selected so as
to maximize (or minimize) an ob8ective function in the form
Max (Min) QM +%K% N +<K< N HHH.N +nKn
(ub8ect to resource constraints in the form of
.%%K% N .%<K< NHHN.%nKn RM -%
.<%K% N .<<K< NHHN .<nKn RM -<
HHHHHHHHHHHHHHHH
.m%K% N .m<K<NHHN.mnKn RM -m
:here +n, .mn, and -m are given constraints.
,epending on the problem, the constraints may also be stated #ith e*ual signs or greater;
than;or ;e*ual to signs.
9.!.1 roduct mi/ decisions using $inear programming
This follo#ing problem is solved using 'raphical method.
&et K and O denote the *uantities of the products K and O. The ob8ective function is
(Profit) given by>
P M DK N FO
(ub8ect to the constraints>
?K N <O <GGHHHHHconstraint for the machine hours
=K N AO <?GHHHHH.constraint for ra# materials.
and K G,O GHHHHnon;negative constraints
The last t#o constraints express that the *uantities of K and O cannot be negative. These
are kno#n as non;negativity constraints, #hich are necessary for all &PPs.7aving
formulated the problem, #e #ill no# solve the problem.
The first constraint can be #ritten as O N <K %GG and similarly the second can be
#ritten as O N G.@K ?G.
%<
&et us plot the above ine*ualities #ith O on the vertical axis and K on the horizontal axis.
-y substituting !G" for one variable and finding the value of other, #e can find points on
both the axis. The lines are the boundaries or the constraints for each resources.
+orresponding to the constraint of ra# materials, the decision variables K and O can take
any of the infinite different values falling belo# the line for this constraint. The other
constraint of machine hours limits or reduces these possibilities further because the
values of K and O have to reconcile #ith not only the machine hour constraint but also
#ith the ra# material constraint. Therefore, K and O can take only those values that
correspond to points, #hich fall #ithin the polygon described by the boundary lines for
all the four constraints.$or this #e shall bring in the ob8ective function; P M DK N FO
,epending upon the values of K and O, the profit margin function P #ill take a number of
values. 2ur ob8ective is to find out the values of K and O at #hich #e can maximize the
profit margin #ithin the constraints.
The Polygon got four corner points #ith its respective co;ordinates of K and O. The co;
ordinates are (G,G), (G, ?G), (@G,G) and (?G,<G)
(ubstitute these values in the ob8ective function and find out the value of P.
.t (?G, <G) the profit is maximum #ith PM ?@G.
(o, K M ?G and O M <G is the optimal product mix for this company.
Chapter :
Aggregate p$anning
:.1Introduction
.ggregate planning is concerned #ith setting up of production rates in terms of
product group or other broad categories for the intermediate term and it precedes the
master production schedule. The main purpose of the aggregate planning is to specify
%=
(G,G)
(?G,<G
(G,?G)
(@G,G)
O
K
the optimal combination of production rate, #orkforce level, and inventory on hand.
Production rate refers to the number of units completed per unit of time. :orkforce
level is the number of #orkers needed for production (production M production rate x
#orkforce level). 0nventory on hand is unused inventory carried over from the
previous period.
The aggregate planning problem can be stated as>
'iven the demand forecast $t for each period t in the planning horizon that extends over T
periods, determine the production level Pt, inventory level 0t, and #orkforce level :t for
periods t M%,<,=H,T that minimize the relevant costs over the planning horizon.
,ifferent capacities used for .ggregate Planning>
%. egular time production capacity
<. (ubcontracting capacity
=. 2vertime capacity
?. 7iring and $iring capacity
2ne can use any one or a combination of the follo#ing strategies for smoothing
fluctuations in demand.
-uilding and utilizing inventory through constant #ork force
5arying the size of the #ork force
2vertime utilization
(ubcontracting
Turning do#n some orders
0f a single strategy is used to meet the demand, then it is called as a pure strategy. 0f a
combination of the above pure strategies is used to meet the demand, then it is called as a
mixed strategy. -ut all these mixed strategies may not be meaningful and practicable to
implement. 2nly a limited combination of these strategies is used in practice.
+osts associated #ith aggregate planning>
%. egular production costs
<. egular payroll costs
=. +ost of changing production rate
?. +ost of carrying inventory
@. (hortage costs (-ackordering cost)
Methods used in aggregate planning>
'raphical method
7euristic method
Tabular method
Transportation model for aggregate planning
&inear programming model for aggregate planning
&inear decision rule
%?
:.! Linear programming mode$ )or aggregate p$anning#
The aggregate planning problem can be formulated as a linear programming model. 0n
this model, all variables can be explicitly included. (o, &P algorithm provides a solution
#ith a mixed strategy such that the total cost of the problem is minimized.
:.!.1 Mode$ "e*e$opment#
.ssumptions>
%. ,emand rate , is deterministic for all future periods
<. The costs of production during regular time are assumed to be piece;#ise linear
=. The costs of changes in production level are approximated by a piece;#ise linear
function.
1otations>
,i P ,emand for the period !i"
Pi ; egular production schedule in period !i"
Mi; maximum regular production capacity in period !i"
Ti P 2vertime production scheduled in period !i"
Oi P Maximum overtime capacity in period !i"
r; regular time production cost per unit
v P overtime production cost per unit
c Pinventory cost per unit per period
h P hiring cost per unit
f P lay;off cost per unit
.i; 1umber of units increased during consecutive periods (0ncrease in production level)
i P 1umber of units decreased during consecutive periods (,ecrease in production level)
:ith these notations, #e are developing a three period problem>
:.!.! Constraints Formu$ation#
egular time production constraints
P%RM M%
P< RM M<
P= RM M=
The above can be #ritten in simple form as Pi RM Mi
2ver;time production constraints>
T% RM O%
T< RM O<
T= RM O=
These can be #ritten as Ti RM Oi
0nventory constraints>
Period %> P% N T% SM ,%
Period <> P% NT% ; ,% N P< N T< SM ,<
%@
Period => P% NT% ; ,% N P< N T<; ,< NP= N T= SM ,=
The above can be #ritten as Pi N Ti .M ,i #here !i" M%,< and =
7iring +onstraints>
Period %> .% SM P%;PG
Period <> .< SM P<;P%
Period => .= SM P=;P<
:here PG;production level in the previous period should be kno#n
The above can be #ritten as .i SM Pi;Pi;%

&ay;off constraints>
Period %> % SM PG;P%
Period <> < SM P%;P
<
Period => = SM P<;P=
(imply i SM Pi;% ;Pi
The ob8ective function #ill be a summation function of different costs.
0t can be #ritten as
Period %> rP% N vT% N h.% N f% N +(P% NT% ; ,% )
Period <> rP< N vT< N h.< N f< N +(P% NT% ; ,% N P< N T<; ,< )
Period => rP= N vT= N h.= Nf= N +(P% NT% ; ,% N P< N T<; ,< NP= N T= P ,=)
.gain on simplification,
!n" i
Min T+ M r Pi N v Ti N h .i N f i N c I (P8 N T8 P ,8)J
!i"M% 8M%
This &P model can be used for aggregate planning by giving values.
ro.$em#
The follo#ing data #ere obtained from .lpha manufacturing company. 6sing the linear
programming, develop a &P model for aggregate planning.
,emand data> ,%M%@G, ,<M <GG and ,=M =GG
+apacity data> egular time capacity> M%M<GG, M<M=GG and M=M%@G
2vertime capacity> O%M@G, O<MD@ and O=M =@
egular time cost !r" M s %G
2vertime cost per unit !v" M s%@
0nventory cost per unit per period !c" Ms @
7iring cost per unit !h" Ms. %@
&ay;off cost per unit !f" Ms %G
6nits produced at !t"MG and PoM %@G
0nitial inventory !0o" M G
1ote> 2ne can solve this &P problem by &01,2 soft#are.
%A
2ne can do#nload the demo version of &01,2 by visiting the #ebsite ###.lindo.com.
0f you cut and paste the program in &01,2 and execute you get the result.
Chapter =
Basic in*entor& concepts and mode$s
A.% 0ntroduction
0n any organization, inventories add an operating flexibility that #ould not other#ise
exist. The inventory can be classified as
%. a# materials
%D
<. :ork;in;process inventory
=. $inished goods inventory
The ra# materials inventory remove dependency bet#een suppliers and plants. The
#ork;in;process inventories remove dependency bet#een machines of a product line. The
finished goods inventory remove dependency bet#een the plant and its customers4
market.
The main functions of inventory are>
%. (moothing out irregularities in supply
<. Minimizing the production cost
=. allo#ing organizations to cope #ith perishable materials
?. Managing the variability in demand
=.1.1 In*entor& decisions#
There are t#o basic inventory decisions managers must take in their #ork.
%. :hen to replenish the inventory of an item
<. 7o# much of an item to order #hen the inventory of the item is to be replenished.
=.! In*entor& Mode$s#
=.!.1 Basic Mode$1EO>4#
0n this model of inventory, orders of e*ual size are placed . Ohe items against an order are
replenished instantaneously and the items are consumed at a constant rate. The purchase
price per unit is the same irrespective of order size.
&et , be the annual demand in units
+o be the ordering cost4order
+c be the carrying cost4unit4year
P be the purchase price of the unit
B be the order size
%E

The number of orders4year M ,4B
.verage inventory M B4<
+ost of ordering4year M (,4B)x +o
+ost of carrying4yearM B4< x +c
Purchase cost4yearM ,xP
The total cost of inventory per year M (,4B)x +o N B4< x +c N ,xP
.s #e #ant to minimize the cost,
,ifferentiating #ith respect to B,
(;,4B
<
)+o N +c4< MG
TTTTTTTTTT
B
U
M V(<+o,4+c) M )conomic 2rder Buantity ()2B)
1umber of orders M ,4 B
U
Time bet#een orders M B
U
4,
This model is called Purchase model .
B
!t"
%F
time
=.!.! 3ariation o) EO> mode$ )or production situation
1Economic .atch ?uantit& mode$4
0f a +ompany manufactures its component #hich is re*uired for its main product, then
the corresponding model of inventory is called WManufacturing modelX . The rate of
consumption of items is uniform throughout the year. The cost of production per unit is
same irrespective of production lot size.
&et , be the annual demand of an item.
!k" be the production rate of the item (1umber of units produced per year)
+o be the cost per set up
+c be the carrying cost per unit per period
P be the cost of production per unit
,uring the period !t%" the item is produced at the rate of !k" units per period and
simultaneously it is consumed at the rate of !," units per period. (o during this period,
the inventory is built at the rate of !k; , units per period. ,uring the period t<, the
production of the item discontinued but the consumption of that item continued. 7ence
the inventory is decreased at the rate of !," units per period during this period.
TTTTTTTTTTTTTTTTT
The formula for this model is )-B M BU M V(<+o,)4+c(%;,4k)
!t
%
" !t
<
"
<G
B
6
.
1
T
0
T
O
T0M)
!t%" MBU4,
!t<" M BUI%;(,4k)J4,
+ycle time M !t%" N !t<"

=.!.( EO> mode$s ,ith price discounts#
:hen items are purchased in bulk, buyers are usually given discount in the purchase
price of goods. This discount may be a step function of purchase *uantity as sho#n
belo#.
Buantity Purchase price
GRMB%Rb% P%
b%RM B<R b< P<
b<RMB=R b= P=
H. H
H. H
bn;%RMB? Pn
The procedure to compute the optimal order size for this situation is given in the steps>
(tep %>
$ind )2B for the nth (last) price break
TTTTTTTTTT
Bn
U
M V<+o,4iPn #here iPnM carrying cost as a percent of Price (i;in percentage)
0f it is greater than or e*ual to bn;% then the optimal order size B
U
M Bn
U
L other#ise go to
step <.
(tep <
$ind )2B for the !n;%"th price break
TTTTTTTTTT
Bn;%
U
M V<+o,4iPn;%
0f it is greater than or e*ual to bn;<, then compute the follo#ing and select least cost
purchase *uantity as the optimal sizeL other#ise go to step =.
(i) Total cost, T+ (Bn;%
U
)
(ii) Total cost, T+ (bn;%)
(tep =
$ind )2B for the !n;<" the break
<%
TTTTTTTTTT
Bn;<
U
M V<+o,4iPn;<
0f it is greater than or e*ual to bn;=, then compute the follo#ing and select least cost
purchase *uantityL other#ise go to step ?.
i. Total cost, T+ (Bn;<
U
)
ii. Total cost T+(bn;<)
iii. Total cost, T+ (bn;%)
(tep ?
+ontinue in this manner until the end of break;ups and select the *uantity #ith lo#est
cost.
Example:
.nnual demand for an item is ?EGG units. 2rdering cost is s. @GG per order. 0nventory
carrying cost is <?Y of the purchase price per unit, per year. The price breaks are as
sho#n belo#.
Buantity Price
(s)
G;%%FF %G
%<GG;%FFF F
<GGG and
above
E
%. $ind the optimal order size
<. 0f the order cost is changed to s.=GG per order, find the optimum order size.
,M?EGG, +oM @GG, i M G.<?, P=M s. E.GG
>(
@
0 A1!/:BB/9CBB451B.!9/C4 01:C1
(ince >(
@
is less than <GGG(b<), go to the next step.
P<Ms F.GG
>!
@
0 A1!/:BB/9CBB451B.!9/D4 019D1
(ince >!
@
is greater than %<GG (b%) find the follo#ing.IT+ MB+c4< N (,4B)x+o NPBJ
TC 1>!
@
4 M Fx?EGG N (@GGx?EGG)4%?F% N (G.<?xFx%?F%)4< M s. ?A,?<G
T+ (b<) M Ex?EGG N (@GGx?EGG)4<GGG N (G.<?xEx<GGG)4< M s. ?%@<G.
<<
The least cost is s. ?%@<G and 2ptimum *uantity is <GGG.
=.!.9 Ser*ice $e*e$ and sa)et& stocE $e*e$
The previous models assumed that demand #as constant and kno#n. 0n the ma8ority of
cases, though, demand is not constant but varies from day to day. (afety stock must
therefore be maintained to provide some level of protection against stock outs. (afety
stock can be defined as the amount of inventory carried in addition to the expected
demand. 0n a normal distribution, this #ould be the mean. $or example, if our average
monthly demand is %GG units and #e expect next month to be the same, if #e carry %<G
units, then #e have <G units of safety stock.
(afety stock can be determined based on many different criteria. . common approach is
for a company to simply state that a certain number of #eeks of supply be kept in safety
stock. 0t is better, though, to use an approach that captures the variability in demand.
$or example, an ob8ective may be something like, !set the safety stock level so that there
#ill only be a @ percent chance of stocking out if demand exceeds =GG units". :e call this
approach to setting safety stock the probability approach.
:hen there are variations in either the demand rate or the lead;time, the possibility of
stock;outs must be dealt #ith. 6nlike the case of constant demand and lead time, it is no
longer completely certain ho# much stock #ill be needed to satisfy demand during lead
time. 5ariations in the demand rate can result in a temporary surge in demand, #hich #ill
drain inventory more *uickly than expected, and variations in delivery times can lengthen
the time a given supply must cover. 0n order to compensate for uncertainties in either
demand rate or lead;time, additional stock must be carried to reduce the risk of stock;out
during the lead;time interval. The follo#ing figure illustrates ho# safety stock can reduce
the risk of a stock;out during lead;time.
<=
(.$)TO (T2+/
2P
&T
Max. Probable demand during &T
)xpected demand during &T
t
1ote that stock;out protection is re*uired only during lead;time. 0f there is a sudden
surge at any point during the cycle, this #ill trigger another order, and once that order is
received, the danger of an imminent stock;out is practically negligible.
0n general, the re;order point (2P) #hen variations exist in either usage or lead;time is>
2P M )xpected demand during lead time N (afety stock
A.<.?.% (ervice level>
-ecause it costs money to hold safety stock, a manger must carefully #eigh the cost of
carrying safety stock against the reduction in stock out risk it provides, since the service
level increases as the risk of stock out decreases.
(ervice level can be defined as the probability that demand #ill not exceed supply during
lead time (i.e., that the amount of stock on hand #ill be sufficient to meet demand).
7ence, a service level of F@Y implies a probability of F@Y that demand #ill not exceed
supply during lead;time. .n e*uivalent statement is that demand #ill be satisfied in F@3
of such instances. 0t does not mean that F@Yof demand #ill be satisfied. The risk of a
stock;out is the compliment of service levelL a customer service level of F@Y implies a
stock;out risk of @Y.
(ervice level M %GG percent ; (tock;out risk
The amount of safety stock that is appropriate for a given situation #ill depend on the
follo#ing factors.
%. The average demand rate and average lead time
<. ,emand and lead time variability
=. The desired service level
=.!.: ro.a.i$it& distri.utions and sa)et& stocE decisions#
A.<.@.% +ase 0;5ariable demand rate, constant lead;time
This model assumes that demand during lead;time is composed of a series of independent
daily demands that can be described by a normal distribution. 0n order to use the model, it
is necessary to kno# the average daily demand rate and its standard deviation. That
information is then used to determine the expected demand and standard deviation of
demand for the lead;time period.
2PM )xpected demand during lead time N (afety stock

M d x&T N z &T (d)
#here
d ;average demand rate
&T;lead time
d Pstandard deviation of demand rate
<?
z Pvalue is taken from standard normal distribution table using the give service level.
A.<.@.< +ase 00;+onstant demand rate, variable lead;time
. variable lead;time can be described by a normal distribution. :hen the lead;time is
normally distributed the expected demand during lead time is e*ual to d&T , and the
standard deviation of demand during lead time is e*ual to d&T.
2PM d&T Nzd&T.
&T.;standard deviation of lead;time
A.<.@.= +ase 000;5ariable demand and 5ariable &ead;time>
The expected demand during lead time is average daily demand multiplied by average
lead time. 0f daily demand is normally distributed and if lead time is also normally
distributed, then total demand during lead;time #ill be normally distributed #ith a mean
e*ual to d(&T)
(tandard deviation of total demand during the lead;time M
<
demand N
<
lead;time
:here demand M(&T) d
lead;time M d&T
2P M d(&T) N z &T
<
d N d
<

<
&T
=.!.= ro.$ems o) in*entor& contro$ o) spare parts
%. Managing spare parts presents some peculiar problems especially in a developing
country like 0ndia. The ma8or problem #ith spare parts is the uncertainty in regard
to demand. The failure of a part is very difficult to predict.
<. .nother problem related to imported machineries. The imported parts are costly
and #e have to invest on spares for emergency.
=. Third problem is #ith technology obsolescence and parts unavailability
The problem of spares has plagued inventory controllers the #orld over. To take an
example, a 8umbo aircraft has approximately %,A?,GGG spare parts #hich are to be
stocked. 7o# much of each is to be stocked at #hich airport or carried on the aircraftC
:hich of these spares among these thousands of items are to be stocked at allC :hat is
the probability of the failure of a specific partC These are the *uestions #hich need to be
ans#ered in order to control inventories.
<@
=.!.F MGSIC+(" s&stems )or sa)et& stocE po$icies
The materials executive handles a large number of items in actual practice, in any
organization. $or example, a refinery has over @G,GGG items consisting of spares,
refractories, consumables, components, chemicals, and other items. The importance of
each item is indeed a ma8or parameter #hile procuring it. (ome of the items may be very
critical, or difficult to procure, or may have to be imported. Moreover, as the number of
items increases, the effective control and time spent on each item diminishes. To
overcome this defect, #e adopt selective control techni*ues to improve the purchase
activity. The fundamental idea behind selective control techni*ues is to put the efforts
#here the results are #orth it. &et us no# discuss the various control mechanisms used
for inventory control based upon different parameters like consumption values, degree of
importance, etc.,
A.<.D.% .-+ +lassification
0t has been found that about %G per cent of the number of items contributes to about DG
per cent of the consumption value kno#n as . category. The middle <G per cent in
number accounts for about <G per cent in annual consumption value, and is kno#n as -
category. The remaining DG per cent accounts for about %G per cent of annual
consumption value. The exact cut;off points for ., -, + #ill vary depending upon the
decision making ability of the purchase executives and hence #ill vary from firm to firm.
A.<.D.< KOQ and $(1 classification
2ne important #ay of finding the usage of materials is according to their movement. $or
this purpose, #e have fast moving, slo# moving and non;moving or $(1 analysis. KOQ
is based on the year;end stores inventory value, K items are the top %G per cent of item,
accounting for DG per cent of stock value, O the middle <G per cent accounting for middle
<G per cent of stock value and Q the remaining items.
A.<.D.= (,)4'2&$4(2(; based on availability
Materials can be classified according to the problems faced in the procurement. :e can
have items that are scarce to obtain, difficult to obtain, easy to obtain; called (,)
analysis. :e also have '2&$, #hich stands for government controlled, ordinarily
available in open market, locally available and foreign supplier. There are also seasonal
changes in availability of products. $or this #e have (2(;seasonal and off;seasonal type.
A.<.D.? 5),4.M45)01; based on criticality
5),;5ital, essential and ,esirable
5)01; vital e*uipment, essential e*uipment, important e*uipment and normal e*uipment
.M;reliability, availability and maintainability
<A
=.!.C MGSIC+(" ana$&sis
The traditional .-+ analysis discusses only the consumption value #ithout considering
the criticality or availability and hence has miserably failed in actual practice. 0f #e adopt
three levels for each of the follo#ing parameters;consumption value, availability and
criticality, #e get ABCH S"E and 3E", then #e land in <D groups and it becomes a
difficult 8ob in practice to follo#;up. 7ence #e advocate t#o levels for each of the three
dimensions; high consumption value4 lo# consumption valueL long lead;time4short lead;
time and critical4non;critical categories. This integrated picture is ca$$edH MGSIC+("
or Mu$ti+unit se$ecti*e in*entor& contro$ three+dimensiona$ approach.
Iigh consumption *a$ue
items
Lo, consumption *a$ue
Long $ead time Short $ead
time
Long $ead time Short $ead
time
C8ITICAL Critica$
LLT
LC3 1
Critica$
SLT
IC3 !
Critica$
LLT
LC3 (
Critica$
SLT
LC3 9
NON+
C8ITICAL
Non+critica$
LLT
IC3 :
Non+critica$
SLT
IC3 =
Non+critica$
LLT
LC3 F
Non+critica$
SLT
LC3 C
A.<.E.% 0nterpretation of M6(0+;=,
7aving grouped all items on the basis of consumption value, criticality and lead;time, let
us no# consider the item in cell =. These are critical from operational vie#, take long
lead;time for procuring, but #ith lo# consumption value. $or this category, the purchase
*uantity must be very large #ith annual ordering. 0n vie# of the lo# consumption value ,
the po#ers of delegation should be such that these are purchased at the lo#est level or
decentralized purchase. The inventory should be very high and the service level #ill be
%GG per cent.
The problem before the purchase manager are related to the items falling in cells % and <,
#hich are critical in nature and have high consumption value. 7e has to exercise strict
control on these items. The ordering is more fre*uent #ith strong follo#;up.
M6(0+;=, is useful to purchase executive, to take decisions on a scientific and practical
basis on the follo#ing aspects>
a) Purchase *uantity
b) $ollo#;up
c) Tighter procedures systems
d) (afety stock
e) .verage inventory
<D
f) Po#ers of delegation
g) 0nformation system
h) .pplication of value engineering and other cost reduction techni*ues
i) ,evelopment of ne# sources
8) 5endor rating
k) $orecasting
l) );channels
0t is a po#erful approach in the direction of cost reduction and application of scientific
management principles in the making of materials as a profit centre.
=.( Materia$s re?uirement p$anning
Materials e*uirement Planning (MP) is a computer;based information system
designed to handle ordering and scheduling of dependent;demand inventories (e.g., ra#
materials, component parts and subassemblies). . production plan for a specified number
of finished products is translated into re*uirements for component parts #orking
back#ard, using lead;time information to determine #hen and ho# much to order.
MP begins #ith a schedule for finished goods that is converted into a schedule of
re*uirements for the subassemblies, component parts and ra# materials that #ill be
needed to produce the finished items in the specified time frame. Thus, MP is designed
to ans#er three *uestions> #hat is needed, ho# much is needed, and #hen is it needed.
The primary inputs of MP are a bill of materials, #hich tells #hat a finished product is
composed of, a master schedule, #hich tells ho# much finished product is desired and
#hen, and an inventory;records file, #hich indicates ho# much inventory is on hand or
on order. This information is processed using various computer programs to determine
the net re*uirements for each period of the planning horizon. 2utputs from the process
include planned;order schedules, order releases, changes, and performance control
reports.
P2-&)M>
<E
Master
Production
(chedule
-ill;of;materials
file
0nventory
records file
MP computer
programs
Changes
Order releases
Planned;order
schedules
E/ception
reports
Planning reports
Performance
control reports
0nventory
Transactions
a)6sing the information given belo#, develop a complete MP material plan.
0t takes % unit of - and < units of + to make one unit of .. .t the
beginning of time period %, the follo#ing information is available>
Item I" >uantit& on Iand Lead Time
. %GG %
- %@G <
+ EG %
The gross re*uirements of item . are <GG units for period ? and <@G units
for period @.
b. 0f the lead time for item . increases by % #eek, and the lead time for item +
increases by % #eek, #hat #ill the revised materials plan look likeC .re there any
problems, #hich need immediate attentionC
MRP for First problem:
<F
. (%)
+ (<)
)1, P2,6+T
(&ead time M %)
+omponent -
(&ead time M <)
+omponent +
(&ead time M %)
- (%)
MRP for
second
problem:
Period 0 1 2 3 4 5
Gross Req. 200 250
Schd.Receip
ts
n h!nd
"n#entor$
100 100 100 100
%et
req&iremen
t
100 250
Pl!nned
order
rele!se
100 250
GR 100 250
SR
'" 150 150 150 50
%R 200
PR 200
GR 200 500
SR
'" (0 (0 (0
%R 120 500
PR 120 500
Period 0 1 2 3 4 5
Gross Req. 200 250
Schd.Receip
ts
n h!nd
"n#entor$
100 100 100 100
%et
req&iremen
t
100 250
Pl!nned
order
rele!se
100 250
GR 100 250
SR
'" 150 150 50
%R 200
PR 200
GR 200 500 500
SR
'" (0 (0 (0
%R 120 500
PR 120 500
=G
.
-
+
.
-
+
$or item +, the planned order release fall into previous period. That means that an order
for material + should be released immediately so that it reaches you in the second #eek.
=.9 Manu)acturing resources p$anning 1M8 II4
0n the early EGs, materials re*uirements planning #ere expanded into a much broader
approach for planning and scheduling the resources of manufacturing firms. This
approach is termed as MP 00. 0t has not replaced MP, nor is it an improved version of
MP. ather, it represents an effort to expand the scope of production resource planning
and to involve other functional areas of the firm in the planning process. . ma8or purpose
of MP 00 is to integrate the primary functions (Production, Marketing and $inance) as
#ell as other functions such as personnel, engineering and purchasing in the planning
process. The MP;00 frame#ork is given in $igure A.%
=%
(trategic and -usiness
planning
(ales 3 2perations
planning
Master production
scheduling
MP 0
,emand
Management
ough;cut
capacity
planning
5endor 4
Material
limitations
Management
decisions
+apacity re*uirement
planning (+P)
5endor re*uirement planning
(5P)
ealisticC ealisticC
(hop floor control
Purchase planning and
control
1o Oes Oes 1o
$igure A.% MP 00 frame#ork
=.9.1 M8 II JStructure and s&stems
Manufacturing esource Planning is a development that seeks to address some of the
shortcomings of MP. 0t includes all of the elements of MP and , it is based around the
-ill of Materials. 0t uses a Master Production (chedule (MP() as its starting point and
uses the three steps of )xplosion, 1etting and 2ffsetting to create the initial schedule.
7o#ever MP 00 includes the follo#ing four ma8or developments from MP>
A.?.%.%. $eedback
MP 00 includes feedback from the shop floor on ho# the #ork has progressed, to all
levels of the schedule so that the next run can be updated on a regular basis. $or this
reason it is sometimes called Z+losed &oop MPZ.
A.?.%.<. esource (cheduling
There is a scheduling capability #ithin the heart of the system that concentrates on the
resources, i.e. the plant and e*uipment re*uired to convert the ra# materials into finished
goods. $or this reason the initials [MPZ no# mean Manufacturing esources Planning.
The advantages of this development are that detailed plans can be put to the shop floor
and can be reported on by operation, #hich offers much tighter control over the plant.
Moreover loading by resource means that capacity is taken into account. The difficulty is
that capacity is only considered after the MP schedule has been prepared. 0t may turn
=<
out that insufficient time #as allo#ed #ithin the MP schedule for the individual
operations to be completed.
A.?.%.=. -atching ules
-atching rules can be incorporatedL indeed they have to be if resource scheduling is to
take place. Most soft#are packages offer a variety of batching rules. Three of the more
important are Z&ot for &otZ, Z)-BZ and ZPart Period +overZ. The term Z&ot for &otZ means
batches that match the orders. Therefore if a company is planning to make %G of Product
. follo#ed by <G of Product -, then the batches throughout the process #ill match this
re*uirement. 0f both . and - re*uire t#o of a certain sub assembly then that #ill be made
in *uantities of <G of . and ?G of -. 0t is the batching implicitly follo#ed in basic MP.
Z)-BZ stands for )conomic -atch Buantity. The batch size is calculated by a formula that
minimises the cost through balancing the set up cost against the cost of stock. \ZPart
Period +overZ means making batches #hose size cover a fixed period of demand. .
policy of making a #eeks re*uirement in one batch is an example.
A.?.%.?. (oft#are extension programmes
. number of other soft#are programmes are included in the MP 00 suite. (ome of these
are further designed to help the scheduling procedure. The most important is ough +ut
+apacity Planning (++P), an initial attempt to match the order load to the capacity
available, by calculating (using a number of simplifying assumptions) the load per
resource. 2verloads are identified and orders can be moved to achieve a balance. This has
been described as ]knocking the mountains (the overloads) into the valleys (periods of
underload)]. 2ther additions are designed to extend the application of the MP 00
package. $or example it may include an option for entering and invoicing sales orders
((ales 2rder Processing). .nother common extension is into stock recording and a third
into cost accounting. . full MP 00 implementation can therefore act as an integrated
database for the company.,ata accuracyThis last development means that the company
must put great emphasis on data accuracy. )rrors in recording in one part of the system
#ill result in problems for all the users. The suppliers of such systems encourage users to
aim for accuracy of bet#een F@Y and FEY.
=.9.! Capacit& re?uirement p$anning#
The output from the MP 0 process that involves manufactured components constitutes a
tentative schedule of factory order releases. +P considers these planned orders, together
#ith #ork already in progress, to evaluate the availability of resources;e*uipment and
manpo#er;in the correct time periods to execute the plan in a more detailed fashion than
is possible #ith rough;cut capacity planning. 0t produces a report that sho#s the capacity
re*uired by #ork center and by time period, to actually execute the MP;0 plan.
+P produces exception reports that sho# #hich #ork center are overloaded and #hich
are under loaded for each time period. +P uses the material plan, the scheduled receipts,
and planned orders in MP;0 as inputs.
Buestions
%. . plant manager of a chemical plant must determine the lot size for a particular
chemical that has a steady demand of =G barrels4day. The production rate is %FG
==
barrels4day, annual demand is %G,@GG barrels, setup cost is ^<GG, annual holding cost
is ^G.<%4barrel, and the plant operates =@G days4year. ,etermine the production order
*uantity.
< . . 7ospital procures its supplies of a material once a year. The total number
procured is <?GG packages in a year. This policy of procuring material once a year is
being *uestioned. The accountants calculate the cost of inventory holding at s. =A
per package per year. 0t is also figured out that the costs of procurement add up to s.
%,<GG per order. :hat inventory policy #ould you advise to this hospitalC
Chapter F
Schedu$ing o) production
F.1 Introduction
(cheduling is considered to be a ma8or task for shop floor productivity improvement. 0t is
the allocation of resources applying the limiting factors of time and cost to perform a
collection of tasks. 0t is the allocation of starts and finish time to each particular order. 0t
is scheduling various 8obs on a set of resources (machines) such that certain performance
measures are optimized.
The single machine;scheduling problem consists of !n" 8obs #ith the same single
operation on each of the 92-(, #hile the flo# shop scheduling problem consists of !n"
8obs #ith !m" operations on each of the 8ob. The 8ob shop;scheduling problem contains
!n" 8obs #ith !m" operations on each of the 8obs, but, in this case, the process se*uence of
the 8obs #ill be different from each other.
F.! Sing$e machine schedu$ing
The basic single machine;scheduling problem is characterized by the follo#ing
conditions.
%. . set of independent, single operation 8obs is available for processing at time zero
<. (et;up time of each 8ob is independent of its position in 8ob se*uence. (o, the set;
up time of each 8ob can be included in its processing time.
=. )ach 8ob is processed till its completion #ithout break
=?
6nder these conditions, one can see one;to;one correspondence bet#een a se*uence of
the !n" 8obs and a total number of se*uences for the single machine problem is n_.
(ome terms used in this scheduling problem are as follo#s>
Processing time: (tj) 0t is the time re*uired to process 8ob !8".
Due date (dj) 0t is the time at #hich the 8ob !8" is to be completed.
Completion time (Cj): 0t is the time at #hich the 8ob !8" is completed in a se*uence.
Ready Time (rj): 0t is the time at #hich 8ob 8 is available for processing. The ready time of
a 8ob is the difference bet#een the arrival time of that 8ob and the time at #hich that 8ob is
taken for processing. 0n the basic model, as per condition %, r8 M G for all 8obs.
Flo time (Fj): 0t is the amount of time 8ob 8 spends in the system. $lo# time is a
measure, #hich indicates the #aiting time of 8obs in a system. This in turn gives some
idea about in;process inventory due to a schedule.
$8 M +8;r8
!ateness (!j): 0t is the amount of time by #hich the completion time of 8ob 8 differs from
the due date.
&8M +8;d8
Tardiness (Tj): Tardiness is the lateness of 8ob 8 if it fails to meet its due date, or zero,
other#ise.
T8 M max `G, +8;d8a
F.!.1 Measures o) er)ormance#
Mean flo# time> %4n $8 #here 8 M%,<,=H.n
Mean Tardiness M %4n T8 #here 8M%,<,=Hn
,ifferent rules used in single machine schedule problem>
%. (hortest processing time ((PT)
<. )arliest dud date (),,)
=. $irst in first out ($0$2)
=@
ro.$em# 9obs> % < = ? @
Processing time> %@ ? @ %? E
0f #e arrange according to (PT, the se*uence #ill be <, =, @, ? and %. The completion
time +8 #ill be given as ?, F, %D, =%, and ?A
The completion time is calculated as cumulative summation.
(e*uence > < = @ ? %
+8 > ? F %D =% ?A
Mean flo# timeM %4@ ( ? N F N %D N =% N ?A) M <%.? hours
F.( F$o, shop schedu$ing 1Kohnson7s a$gorithm4
0n flo# shop scheduling problem, there are n 8obsL each re*uires processing on m
machines. The order in #hich the machines are re*uired to process a 8ob is called process
se*uence of that 8ob. The process se*uence of all the 8obs is the same. -ut the processing
times for various 8obs on a machine may differ.
$or a problem #ith < machines and n 8obs, 9ohnson had developed a polynomial
algorithm to get optimal solution.
9ob (i) % < = ? ;;;;; ;;;;;;; n
M% T%% T<% T=% T?% HH HH Tn%
M< T%< T<< T=< T?< HHH. HH.. Tn<
(teps 0nvolved>
%. $ind the minimum among various Ti% and Ti<
<. 0f the minimum processing time falls on M%, place the associated 8ob in the first
available position in se*uence. -ut if the minimum processing time re*uires
machine <, place the associated 8ob in the last available position in se*uence.
=. emove the assigned 8ob from consideration and return to step % until all
positions in se*uence are filled.
)xample>
=A
9ob (i) M% M<
. @ ?
- < =
+ %= %?
, %G %
) E F
$ %< %%
(olution>
(tage 6nscheduled 8obs Minimum .ssignment Partial se*uence
% .,-,+,,,),$ T?< ;% ,M A
T7
P2(0T021
KKKKK,
< .,-,+,),$ T<%;< -M%
(T
P2(0T021
-KKKK,
= .,+,),$ T%<;? .M&.(T -6T
21)
-KKK.,
? +,),$ T@%;E )M()+21, -)KK.,
@ +,$ TA<;%% $M$26T7 -)K$.,
A + T=%;%= +MT70, -)+$.,
The optimal se*uence is -)+$.,
The calculation of completion time and idle time are as follo#s>
9ob M%;Time in M%;Time out M<;Time in M<;Timeout 0dle time on M<
- G < < @ <
) < %G %G %F @
=D
+ %G <= <= =D ?
$ <= =@ =D ?E G
. =@ ?G ?E @< G
, ?G @G @< @= G
The make span for this schedule is @=.
Buestions
%. The follo#ing table provides details about 8obs, their processing times and due days.
6se $+$(, ),,, and (PT se*uencing rules and evaluate in terms of mean flo# time,
average tardiness and number of tardy 8obs.
9ob Processing time (7ours) ,ue time (7ours)
. A F
- = =
+ < ?
, D %?
) ? D
$ E %=
<. (chedule the follo#ing six 8obs through t#o machines in se*uence to minimize the
flo# time using 9ohnson"s ruleL
9ob 2peration time;M% 2peration time;M<
. @ <
- %E %@
+ % F
, %= %%
) %D =
$ %E D
=E
Chapter C
E8T5CM in ro<ect management
C.1 rogram e*a$uation and re*ie, techni?ue 1E8T4 and Critica$ path
method 1CM4
-oth methods sho# precedence relationships explicitly. .lthough the t#o methods #ere
developed independently during the fifties, they are surprisingly similar. -oth methods,
P)T and +PM, use a graphic representation of a pro8ect that it is called ]Pro8ect
1et#ork] or ]+PM diagram], and it is used to portray graphically the interrelationships
of the elements of a pro8ect and to sho# the order in #hich the activities must be
performed.
C.1.1 S&m.o$s and notations used in the net,orE
0n order to represent a pro8ect net#ork, t#o basic elements are used>
. cycle, called ]node], represents an event. .n event describes a checkpoint. 0t does not
symbolize the performance of #ork, but it represents the point in time in #hich the event
is accomplished.
.n arro#, called ]arc], represents an activity.
The net#ork #ill try to reflect all the relationships bet#een the activities.
T#o simple rules govern the construction of a pro8ect net#ork>
.. )ach activity must be represented by only one directed arc or arro#.
-. 1o t#o activities can begin and end on the same t#o nodes or cycles.
.nother element to represent a pro8ect net#ork is a ]dummy activity]. To explain it, #e
#ill consider the follo#ing example>
=F
The temptation is to represent these relationships as>
-ut then #e have broken the second rule earlier mentioned. To sho# that activities . and
- precede +, #hereas activity - precedes activity ,, #e use a dummy activity as sho#n
in the figure>
To construct a pro8ect net#ork, first of all, #e need a list of activities sho#ing the
precedence relationships bet#een the different activities involved as sho#n in the
follo#ing Table E.%
?G
Table E.%. Precedence relationships bet#een activities
C.1.1 8OKECT NET%O8L
-ecause each activity must have a uni*ue pair of starting and ending nodes, #e must use
a dummy activity to dra# the first four activities, as sho#n in the figure. +onstructing a
pro8ect net#ork is a trial;and;error process. 0t usually takes t#o or three attempts to
produce a neatly constructed net#ork. .fter constructing the net#ork, the duration of
each activity should be sho#n in parenthesis. -ut, #hat is this forC :ith this
representation #e can determine the minimum completion time for the pro8ect. :e do this
by starting at the originating event of the net#ork (node %) and determining the earliest
?%
time #e can start an activity, given the activities that precede it and assuming that all the
activities start as soon as possible and are completed as soon as possible. $or example for
the first one, it #ould be>
:here TMG is the )arliest (tart Time ()() for activity . and TM% is the &atest (tart Time
()$) for activity .. +ontinuing this process results in the net#ork>
E.%.%.% P29)+T 1)T:2/ :0T7 ).&0)(T T0M)
1otice that in the case of activity , for example, it only starts after both precedence
activities - and + are completed. 0f everything goes as planned, the pro8ect #ill take %@.@
months to complete. 7o#ever, every activity needs to start as early as possible for the
pro8ect to be completed in %@.@ months. :e can use a similar process to determine #hich
activities #e can delay, and by ho# much, #ithout increasing the completion time of the
pro8ect. To calculate this, #e can define the ]&atest $inish Time] (&$), and the ]latest
(tart Time] (&() for each activity, for example>
+ontinuing #ith this process #e can obtain
E.%.%.< P29)+T 1)T:2/ :0T7 ).&0)(T .1, &.T)(T T0M)(
1o# #e have a pro8ect net#ork #ith the earliest and the latest start and finish times,
#here>
?<
Then 0 can calculate ho# much 0 can delay an activity. That is the ](lack Time.] To
determine it, #e can use either or t#o e*uations>
(lack Time> &(;)(
(lack Time> &$;)$
The slack represents ho# long #e can delay the activity #ithout delaying the entire
pro8ect. The activities that have zero slack lie on a path through the net#ork. This path is
called the ]+ritical Path,] and the activities are called ]+ritical .ctivities.] 0f you delay
these activities, you #ill delay the entire pro8ect. )very pro8ect has at least one critical
path, but there can be more than one. .nother procedure to determine the critical path is
8ust noticing #hich is the largest path through the net#ork, in this case .;-;,;';7;0;/;
&.
C.1.! E8T5CM# "IFFE8ENCES
-oth tools lead to the same end> a critial path and critical activities #ith slack time e*ual
to zero. The differences bet#een these tools come from ho# they treat the activity time.
P)T treats activity time as a random variable #hereas +PM re*uires a single
deterministic time value for each activity. .nother difference is that P)T focuses
exclusively on the time variable #hereas +PM includes the analysis of the Time4+ost
Trade;off.
:e have a high degree of uncertainty in regard to the completion time of many activities.
0t makes sense in the real #orld that you do not really kno# ho# long a particular activity
#ill take, specially talking about certain activities such as research and development. 0n
this case, #e can look at the pro8ect completion time in a probabilistic fashion and for
each activity #e can define>
a. 2ptimistic time estimate> an estimate of the minimum time an activity #ill
re*uire.
b. Most likely time estimate> an estimate of the normal time an activity #ill re*uire.
c. Pessimistic time estimate> an estimate of the maximum time an activity #ill
re*uire.
These three estimates are considered to be related in the form of unimodal probability
distribution> m. :hat #e need in any case is a specific duration for each activity taking
into consideration these three estimates. This can be possible calculating the expected or
mean activity time for each activity as
?=
:ith the expected time for each activity #e can determine #hich the critical path.. 6sing
three assumptions, #e can conclude that pro8ect completion time or critical path
completion time has a normal distribution. 6sing this, #e can determine probabilities,
using completion time as a normal random variable, mean and standard deviation.
Pro"lem
%. . pro8ect consisting of %< district activities is to be analyzed by using P)T. The
follo#ing information is given (Time estimates are in days)
.ctivity Predecessor
activity
2ptimistic time
(ta)
1ormal tie (tm) Pessimistic time
(tb)
. ; < < <
- ; % = D
+ . ? D E
, . = @ D
) - < A F
$ - @ = %%
' +, , = A E
7 ) < A F
0 +, , = @ E
9 ', 7 % = ?
/ $ ? E %%
& 9, / < @ D
(0) ,ra# the P)T net#ork.
(00) Mark the critical path in bold #ay..
(000) :ithin ho# many days can the pro8ect be expected to be completed #ith FF
per cent chanceC
(. normal curve #ith its values are attached)
.ctivity Predecessor
activity
2ptimistic
time (ta)
1ormal
tie (tm)
Pessimistic
time (tb)
!te"M(taN?tmNtb)4A 5arM
I(tb;ta)4Aa
<
. ; < < < < G
- ; % = D =.== %
+ . ? D E A.AA G.???
, . = @ D @ G.A
??
) - < A F @.E= %.=A
$ - @ = %% ?.AA %
' +, , = A E @.E= G.AF
7 ) < A F @.E= %.=A
0 +, , = @ E @.%D G.AF
9 ', 7 % = ? <.E= G.<@
/ $ ? E %% D.E= %.=A
& 9, / < @ D ?.E= G.AF
,
. + 0
%
- ) 7 9 &
$ /
?@
G
%
<
=
?
@
A
D
E
F
%
=
<
< D
F
<?
<?
%
; for#ard calculation
; reverse calculation
; nodes
Three critical paths are
%. .;+;0
<. .;+;';9;&
=. -;);7;9;&
2ut of these critical paths, .+0 total variance MGNG???NG.AFM %.%=? (refer 5ariance
column of the table)
.+'7& variance MGNG.???NG.AFNG.<@NG.AFM<.GD?
-)79& variance M %N%=AN%.=ANG.<@NG.AF M?.AA
2ut of these values, .+0 is minimum. Oou can take that as critical path.
Pro8ect completion time>
QM(,ue date; )xpexted time of pro8ect completion)4(*uare root of Pro8ect variance
$pr FFY chance the Q value is <.==
$or critical path one, .+7, <.==Mdue date P<?4%.%=, ,ue dateM<A.?@
$or critical path .+'9&, due dateM <D.=@
$or critical path -)79&, due date M <F
The pro8ect is expected to be completed #ithin <D days. (&o#est of above three)
>uestions
%. . pro8ect consisting of %< distinct activities is to be analyzed by using P)T. The
follo#ing information is given (Time estimates are in days)
.ctivity Predecessor
activity
2ptimistic time
(ta)
1ormal tie (tm) Pessimistic time
(tb)
?A
G
<
?
@
A
D
E
F
G
G
=
=
F
F
%@
%@
F
F
E
%G
%E
%E
. ; < < <
- ; % = D
+ . ? D E
, . = @ D
) - < A F
$ - @ = %%
' +, , = A E
7 ) < A F
0 +, , = @ E
9 ', 7 % = ?
/ $ ? E %%
& 9, / < @ D
a. ,ra# the P)T net#ork.
b. Mark the critical path in bold #ay
Chapter D
>ua$it& and statistica$ process contro$
D.1 Introduction
The concept of *uality has been #ith us since the beginning of time. .s early as the
creation of the #orld described in the -ible in 'enesis, 'od pronounced his creation
]good];; e.g., acceptable *uality. .rtisansZ and craftsmenZs skills and the *uality of their
#ork are described throughout history. Typically the *uality intrinsic to their products
#as described by some attribute of the products such as strength, beauty or finish.
7o#ever, it #as not until the advent of the mass production of products that the
reproducibility of the size or shape of a product became a *uality issue.
Buality, particularly the dimensions of component parts, became a very serious issue
because no longer #ere the parts hand;built and individually fitted until the product
#orked. 1o#, the mass;produced part had to function properly in every product built.
Buality #as obtained by inspecting each part and passing only those that met
specifications. This #as true until %F=% #hen :alter (he#hart, a statistician at the
7a#thorne plant at :estern )lectric, published his book Economic Control o# $uality o#
%anu#actured Product (5an 1ostrand, %F=%). This book is the foundation of modern
statistical process control ((P+) and provides the basis for the philosophy of total *uality
?D
management or continuous process improvement for improving processes. :ith
statistical process control, the process is monitored through sampling. +onsidering the
results of the sample, ad8ustments are made to the process before the process is able to
produce defective parts.
D.! rocesses and rocess 3aria.i$it&
The concept of process variability forms the heart of statistical process control. $or
example, if a basketball player shot free thro#s in practice, and the player shot %GG free
thro#s every day, the player #ould not get exactly the same number of baskets each day.
(ome days the player #ould get E? of %GG, some days AD of %GG, some days DD of %GG,
and so on. .ll processes have this kind of variation or variability.
This process variation can be partitioned into t#o components. 1atural process variation,
fre*uently called common cause or system variation, is the naturally occurring fluctuation
or variation inherent in all processes. 0n the case of the basketball player, this variation
#ould fluctuate around the playerZs long;run percentage of free thro#s made. (pecial
cause variation is typically caused by some problem or extraordinary occurrence in the
system. 0n the case of the basketball player, a hand in8ury might cause the player to miss a
larger than usual number of free thro#s on a particular day.
D.!.1 Statistica$ rocess Contro$
(he#hartZs discovery statistical process control or (P+, is a methodology for charting the
process and *uickly determining #hen a process is ]out of control] (e.g., a special cause
variation is present because something unusual is occurring in the process). The process
is then investigated to determine the root cause of the ]out of control] condition. :hen
the root cause of the problem is determined, a strategy is identified to correct it. The
investigation and subse*uent correction strategy is fre*uently a team process and one or
more of the TBM process improvement tools are used to identify the root cause. 7ence,
there is emphasis on team#ork and training in process improvement methodology.
0t is managementZs responsibility to reduce common cause or system variation as #ell as
special cause variation. This is done through process improvement techni*ues, investing
in ne# technology, or reengineering the process to have fe#er steps and therefore less
variation. Management #ants as little total variation in a process as possible;;both
common cause and special cause variation. educed variation makes the process more
predictable #ith process output closer to the desired or nominal value. The desire for
absolutely minimal variation mandates #orking to#ard the goal of reduced process
variation.
?E
$igure F.% Process in statistical control
The process in $igure F.% is in apparent statistical control. 1otice that all points lie #ithin
the upper control limits (6+&) and the lo#er control limits (&+&). This process exhibits
only common cause variation.
$igure F.< Process out of statistical control
The process in $igure F.< is out of statistical control. 1otice that a single point can be
found outside the control limits (above them). This means that a source of special cause
variation is present. The likelihood of this happening by chance is only about % in %,GGG.
This small probability means that #hen a point is found outside the control limits that it is
very likely that a source of special cause variation is present and should be isolated and
dealt #ith. 7aving a point outside the control limits is the most easily detectable out;of;
control condition.
$igure F.= Process in typical cycle in (P+
?F
The $igure F.= illustrates the typical cycle in (P+. $irst, the process is highly variable
and out of statistical control. (econd, as special causes of variation are found, the process
comes into statistical control. $inally, through process improvement, variation is reduced.
This is seen from the narro#ing of the control limits. )liminating special cause variation
keeps the process in controlL process improvement reduces the process variation and
moves the control limits in to#ard the centerline of the process.

D.!.! Samp$ing and Contro$ Charts
(everal issues are important #hen selecting sample for control chart purposes. They
include>
sample (subgroup) size considerations
sampling fre*uency
collecting samples
. ma8or goal #hen selecting a sample from a process is to select the sample (subgroup)
in such a #ay that the variation #ithin the subgroup is attributable only to the random
variation inherent in the process or common cause variation. The idea is that the sample
should be chosen in such a manner that the chances are maximized to have each element
in the sample be alike and sub8ect only to common cause variation. The spacing of the
samples (subgroups) is arranged so that i# special cause variation is present, the control
chart can identify its presence.
F.<.<.% +onsiderations in ,etermining (ubgroup ((ample) (ize
(ubgroups should be small enough to be economically and practically feasible.
The time and effort it takes to collect and measure samples #eighs heavily here.
(ubgroups should be large enough to allo# the central limit theorem to cause the
distribution of sample means to be normally distributed. 0 many cases the process
measurements are normal, or close to normal. 0n a fe# cases they may not be
normal. :e kno# from the central limit theorem that the larger the sample size,
the more likely it is that the distribution of sample means #ill follo# the normal
distribution. $rom a practical perspective, this is true for most subgroup sizes of ?
or more.
&arger subgroups also are needed to provide good sensitivity in detecting out of
control conditions. The larger the subgroup size, the more likely it is that a shift in
a process mean #ould be detected.
.s mentioned previously, subgroups should be selected so that they are sub8ect
21&O to common cause variation. 0f subgroups are allo#ed to get very large, it is
possible that special cause variation can be mixed #ith common cause. This effect
#ill reduce the sensitivity of the control chart in detecting shifts in the process
characteristic of interest.
@G
:hen all of the above considerations is taken into account, typically a subgroup size of
bet#een ? and A is selected. $ive is the most commonly used subgroup size (this is due to
the historic fact that since @ is half of %G, many computations #ith @ can be done
mentally;;#ith calculators and computers, this is probably no longer an important
consideration.
F.<.<.< +onsiderations in +ollecting (amples
Typically, #e #ant measurement #ithin a subgroup to be taken as close to the same time
as possible to reduce the change that special cause variation is present #ithin the
subgroup. Thus it is common that consecutive samples from a process are taken. . period
of time elapses, and another subgroup sample is collected consecutively.
The spacing bet#een the subgroups shouldnZt be exactly uniform. 0t is not a good idea to
take samples )K.+T&O every hour or at )K.+T&O the same time each day. . certain
amount of randomness in the interval bet#een samples is good because it tends to
minimize the effect of shift changes, tool #ear, tool changes, etc. 0f the rule is to take
samples hourly, a better plan might be to take them hourly, but vary the time randomly
#ithin %G minutes of the hour interval.
F.<.<.= The $re*uency of (ample +ollecting
The bottom line is that samples must be collected fre*uently enough to be useful in
identifying and solving problems. 0n many cases in industry, samples are collected too
infre*uently. The follo#ing should be considered>
Process Stability;;0f a process has not been analyzed using control charts before
and exhibits erratic behavior, sampling should be more fre*uent to increase the
opportunities for process improvement. 0n this case, fre*uently .&& parts are
sampled, measured, and grouped serially into groups of @, for example, and then
charted. The fre*uency bet#een these samples of @ is reduced as the process
becomes more stable.
Frequency of Process Events;;0f a process has may things happening in it,
material changes, tool changes, process ad8ustments, etc., sampling should take
place after these potential special causes so they can be detected. :hen many
special events are taking place in a process, each shift, taking t#o samples
(subgroups) per shift #ill be of little benefit.
Sampling Cost;;T#o considerations can occur. The time involved in taking the
sample is one factor and if the *uality characteristic can be observed only through
destructive testing, the loss or output can be a significant cost. These costs must
be #eighed #hen determining the fre*uency of sampling. . much more usual
condition is that the sampling cost is deemed too high and the fre*uency of
sampling is reduced to the level that the charts derived have no value. 0n this case,
a great deal is spent on sampling #ith no value derived from the charts and .&&
of the expenditures are #asted. Thus, if the process is to be sampled, the samples
@%
should be taken fre*uently enough that the resulting charts are of value. )lse,
charting these processes should simply be abandoned.
F.<.<.? The Problems of (tratification and Mixing
Stratification occurs #hen the output of several parallel (and assumed identical)
processes into a single sample for charting the combined process. Typically a single
sample is taken from each machine and included in the subgroup. 0f a problem develops
in the process for one of the machines, it is very difficult to detect because the sample
from the ]problem] machine is grouped #ith other samples from ]normal] machines.
:hat is plotted as common cause variation is really common cause variation plus the
slight differences bet#een the process means of the individual machines. Typically,
stratification is detected #hen large numbers of points lie in Qone + of an chart. 0t looks
like the process is in super control #hen, in fact, the control limits are 8ust calculated too
#ide. The solution to stratification, obviously, is to chart each machine separately.
+ontrol charts are applicable to one and only one process at a time.
Mixing is similar to stratification, except the output of several parallel machines is mixed
and the periodic sample is dra#n from the mixture. (imilar to stratification, mixing #ill
mask problems in individual machines and #ill make isolation of the problem difficult.
Mixing tends to produce an appearance on the control chart #here points tend to lie near
the control limits that they really should be. The more dissimilar the machines, the more
pronounced this phenomina #ill be. $re*uently mixtures come from processes such as
multispindle scre# machines, multicavity molds. The solution to mixing, obviously, is to
chart each machine or mold separately. +ontrol charts are applicable to one and only one
process at a time.
F.= rocess Capa.i$it& Ana$&sis
The capability of a process is defined as the inherent variability of a process in the
absence of any undesirable special causesL the smallest variability of #hich the process is
capable #ith variability due solely to common causes.
Typically, processes follo# the normal probability distribution. :hen this is true, a high
percentage of the process measurements fall bet#een of the process mean or center.
That is, approximately .<DY of the measurements #ould naturally fall outside the
limits and the balance of them (approximately FF.D=Y) #ould be #ithin the limits.
(ince the process limits extend from to , the total spread amounts to about
total variation. 0f #e compare process spread #ith specification spread, #e typically have
one of three situations>
D.(.1 Case I +A Iigh$& Capa.$e rocess
The process spread is ,e$$ ,ithin the speci)ication spread
@<
Ab M NGSL+LSLO
:hen processes are capable, #e have an attractive situation for several reasons>
:e could tighten our specification limits and claim our product is more uniform or
consistent than our competitors. :e can rightfully claim that the customer should
experience less difficulty, less re#ork, more reliability, etc. This should translate into
higher profits.
D.(.! Case II +A Bare$& Capa.$e rocess
The rocess Spread Kust A.out Matches
Ab MI6(&;&(&J


:hen a process spread is 8ust about e*ual to the specification spread, the process is
capable of meeting specifications, but barely so. This suggests that if the process mean
moves to the right or to the left 8ust a little bit, a significant amount of the output #ill
exceed one of the specification limits. The process must be #atched closely to detect
shifts from the mean. +ontrol charts are excellent tools to do this.
D.(.( Case III+A Not Capa.$e rocess
@=
The rocess Spread Is %ithin
Ab SI6(&;&(&J

:hen the process spread is greater than the specification spread, a process is not capable
of meeting specifications regardless of #here the process mean or center is located. This
is indeed a sorry situation. $re*uently this happens, and the people responsible are not
even a#are of it. 2ver ad8ustment of the process is one conse*uence, resulting in even
greater variability. .lternatives include>
+hanging the process to a more reliable technology or studying the process
carefully in an attempt to reduce process variability.
&ive #ith the current process and sort %GGY of the output.
e;center the process to minimize the total losses outside the spec limits
(hut do#n the process and get out of that business.

D.(.9 Steps in "etermining rocess Capa.i$it&
,etermine if (pecifications are +urrently -eing Met
%. +ollect at least %GG random samples from the process
<. +alculate the sample mean and standard deviation to estimate the true mean and
standard deviation of the process.
=. +reate a fre*uency distribution for the data and determine if it is close to being
normally distributed. 0f it is continueL if not, get the help of a statistician to
transform the data or to find an alternative model.
?. Plot the 6(& and &(& on the fre*uency distribution.
@. 0f part of the histogram is outside the specification limits, consider ad8usting the
mean to center the process.
A. 0f the histogram indicates that the process spread is greater than the specification
spread, the process might not be capable.
@?
,etermine the 0nherent 5ariability 6sing an r;+hart
%. 'et at least ?G rational subgroups of sample size, preferably at least ? or @.
<. +alculate the ranges for each subgroup, the average range, rbar, and the control
limits for an r;chart. Plot the data.
=. ,iscard any ranges outside the 6+& 21&O if the undesirable special cause is
identifiable and can be removed from the processL other#ise include the offending
range(s).
?. ecalculate the average range, rbar, and the control limits.
@. epeat the last t#o steps until all ranges are in statistical control.
A. )stimate the process standard deviation, , #ith the formula .
D. 6sing the midpoint of the specifications as the process mean, assuming normality,
dra# a normal curve, and estimate the percentage meeting specifications. This
step assumes that the mean of the process can be ad8usted or recentered.
E. 0f the normal curve sho#s that the specifications can be met, determine #hat the
specifications are not being met. +reate an xbar chart from the same subgroup
data and hunt for clues. 0t may be as simple as recentering the process. Perhaps
special causes are present that can be removed.
F. 0f the specifications cannot be met, consider changing the process by
improvement, living #ith it and sorting %GGY of the output, centering the mean,
or dropping the product totally.
%G. (et up a system of xbar;r charts to create future process improvements and
process control.

D.9 Capa.i$it& Indices
+apability indices are simplified measures to *uickly describe the relationship bet#een
the variability of a process and the spread of the specification limits. &ike many
simplified measures, such as the grades ., -, +, ,, and $ in school, capability indices do
not completely describe #hat is happening #ith a process. They are useful #hen the
assumptions for using them are met to compare the capabilities of processes.
D.9.1 The Capa.i$it& Inde/ + C
p
The e*uation for the simplest capability index, +p, is the ratio of the specification spread
to the process spread, the latter represented by six standard deviations or Ab.
+p assumes that the normal distribution is the correct model for the process. +p can be
highly inaccurate and lead to misleading conclusions about the process #hen the process
data does not follo# the normal distribution.
@@
2ccasionally the inverse of the capability index +p, the capability ratio CR is used to
describe the percentage of the specification spread that is occupied or used by the process
spread.
+p can be translated directly to the percentage or proportion of nonconforming product
outside specifications. :hen +p M%.GG, approximately .<DY of the parts are outside the
specification limits (assuming the process is centered on the midpoint bet#een the
specification limits) because the specification limits closely match the process 6+& and
&+&. :e say this is about <DGG parts per million (ppm) nonconforming.
:hen +p M%.==, approximately .GGA?Y of the parts are outside the specification limits
(assuming the process is centered on the midpoint bet#een the specification limits). :e
say this is about A? parts per million (ppm) nonconforming. 0n this case, #e #ould be
looking at normal curve areas beyond %.==x=b MN or ; ?b from the center.
:hen +p M%.AD, approximately .GGGG@DY of the parts are outside the specification limits
(assuming the process is centered on the midpoint bet#een the specification limits). :e
say this is about .A parts per million (ppm) nonconforming. 0n this case, #e #ould be
looking at normal curve areas beyond %.ADx=b MM N or ; @b from the center of the
normal distribution. emember that the capability index +p ignores the mean or target of
the process. 0f the process mean lined up exactly #ith one of the specification limits, half
the output #ould be nonconforming regardless of #hat the value of +p #as. Thus, +p is a
measure of potential to meet specification but says little about current per#ormance in
doing so.
D.9.! The Capa.i$it& Inde/ +C
pE
The ma8or #eakness in +p #as the fact that fe#, if any processes remain centered on the
process mean. Thus, to get a better measure of the current performance of a process, one
must consider #here the process mean is located relative to the specification limits. The
index +pk #as created to do exactly this. :ith +pk , the location of the process center
compared to the 6(& and &(& is included in the computations and a #orst case scenario
is computed in #hich +p is computed for the closest specification limit to the process
mean.
:e have the follo#ing situation. The process standard deviation is b M.E #ith a 6(&M<?,
&(&M%E, and the process mean cM<<.
@A

M min`G.E= and %.ADaM G.E=
0f this processZ mean #as exactly centered bet#een the specification limits,
.

D.9.( The Capa.i$it& Inde/ + C
pm
+pm is called the Taguchi capability index after the 9apanese *uality guru, 'enichi Taguchi
#hose #ork on the Taguchi &oss $unction stressed the economic loss incurred as
processes departed from target values. This index #as developed in the late %FEGZs and
takes into account the proximity of the process mean to a designated target, T.
:hen the process mean is centered bet#een the specification limits and the process mean
is on the target, T, Cp & Cp' & Cpm(
:hen a process mean departs from the target value T, there is a substantive affect on the
capability index. 0n the Cp' example above, if the target value #ere TM<%, +pm #ould be
calculated as>
M%.<E%
0n this case, the Taguchi capability index is some#hat more liberal than Cp'(

@D
D.: Motoro$aPs Si/ Sigma >ua$it&
0n %FEE, the Motorola +orporation #as the #inner of the Malcolm -aldrige 1ational
Buality .#ard. Motorola bases much of its *uality effort on #hat its calls its ]A;(igma]
Program. The goal of this program #as to reduce the variation in every process to such an
extent that a spread of %< (A on each side of the mean) fits #ithin the process
specification limits. Motorola allocates %.@ on either side of the process mean for
shifting of the mean, leaving ?.@ bet#een this safety zone and the respective process
specification limit.


Thus, even if the process mean strays as much as %.@ from the process center, a full ?.@
remains. This insures a #orst case scenerio of =.? ppm nonconforming on each side of
the distribution (A.E ppm total) and a best case scenerio of % nonconforming part per
.i$$ion (ppb) for each side of the distribution (< ppb total). 0f the process mean #ere
centered, this #ould translate into a CpM<.GG.
Motorola has made significant progress to#ard this goal across most processes, including
many office and business processes as #ell.
D.= XBA8 and 8 Charts
Theoretica$ Contro$ Limits )or XBA8 Charts
.lthough theoretically possible, since #e do not kno# either the population process mean
or standard deviation, these formulas cannot be used directly and both must be estimated
@E
from the process itself. $irst the chart is constructed. 0f the chart validates that the
process variation is in statistical control, the K-. chart is constructed.
D.=.1 Steps in Constructing an 8 Chart
%. (elect k successive subgroups #here k is at least <G, in #hich there are n
measurements in each subgroup. Typically n is bet#een % and F. =, ?, or @
measurements per subgroup is *uite common.
<. $ind the range of each subgroup (i) #here (i)Mbiggest value ; smallest value
for each subgroup i.
=. $ind the centerline for the chart, denoted by
?. $ind the 6+& and &+& #ith the follo#ing formulas> 6+&M ,(?)-. and
&+&M,(=)-. #ith ,(=) and ,(?) can be found in the follo#ing table>
Table of ,(=) and ,(?)
n ,(=) ,(?) n ,(=) ,(?)
< G =.<AD A G <.GG?
= G <.@D? D .GDA %.F<?
? G <.<E< E .%=A %.EA?
@ G <.%%? F .%E? %.E%A
@. Plot the subgroup data and determine if the process is in statistical control. 0f not,
determine the reason for the assignable cause, eliminate it, and the subgroup(s)
and repeat the previous = steps. ,o 12T eliminate subgroups #ith points out of
range for #hich assignable causes cannot be found.
A. 2nce the chart is in a state of statistical control and the centerline -. can be
considered a reliable estimate of the range, the process standard deviation can be
estimated using>
d(<) can be found in the follo#ing table>
n d(<) n d(<)
< %.%<E A <.@=?
= %.AF= D <.DG?
? <.G@F E <.E?D
@ <.=<A F <.FDG
D.=.! Steps in Constructing the XBA8 Chart
@F
%. $ind the mean of each subgroup K-.(%), K-.(<), K-.(=)... K-.(k) and
the grand mean of all subgroups using>
<. $ind the 6+& and &+& using the follo#ing e*uations>
.(<) can be found in the follo#ing table>
n .(<) n .(<)
< %.EEG A .?E=
= %.G<= D .?%F
? .D<F E .=D=
@ .@DD F .==D
=. Plot the &+&, 6+&, centerline, and subgroup means
?. 0nterpret the data using the follo#ing guidelines to determine if the process is in
control>
a. one pount outside the = sigma control limits
b. eight successive points on the same side of the centerline
c. six successive points that increase or decrease
d. t#o out of three points that are on the same side of the centerline,
both at a distance exceeding < sigmas from the centerline
e. four out of five points that are on the same side of the centerline,
four at a distance exceeding % sigma from the centerline
f. using an average run length (.&) for determining process anomolies
E/amp$e#
The follo#ing data consists of <G sets of three measurements of the diameter of an engine
shaft.
n measQ1 measQ! measQ( 8ange XBA8
% <.GGGG %.FFFE <.GGG< G.GGG? <.GGGG
< %.FFFE <.GGG= <.GGG< G.GGG@ <.GGG%
= %.FFFE <.GGG% <.GGG@ G.GGGD <.GGG%
? %.FFFD <.GGGG <.GGG? G.GGGD <.GGGG
@ <.GGG= <.GGG= <.GGG< G.GGG% <.GGG=
A <.GGG? <.GGG= <.GGGG G.GGG? <.GGG<
AG
D %.FFFE %.FFFE %.FFFE G.GGGG %.FFFE
E <.GGGG <.GGG% <.GGG% G.GGG% <.GGG%
F <.GGG@ <.GGGG %.FFFF G.GGGA <.GGG%
%G %.FFF@ %.FFFE <.GGG% G.GGGA %.FFFE
%% <.GGG< %.FFFF <.GGG% G.GGG= <.GGG%
%< <.GGG< %.FFFE <.GGG@ G.GGGD <.GGG<
%= <.GGGG <.GGG% %.FFFE G.GGG= <.GGGG
%? <.GGGG <.GGG< <.GGG? G.GGG? <.GGG<
%@ %.FFF? <.GGG% %.FFFA G.GGGD %.FFFD
%A %.FFFF <.GGG= %.FFF= G.GG%G %.FFFE
%D <.GGG< %.FFFE <.GGG? G.GGGA <.GGG%
%E <.GGGG <.GGG% <.GGG% G.GGG% <.GGG%
%F %.FFFD %.FFF? %.FFFE G.GGG? %.FFFA
<G <.GGG= <.GGGD %.FFFF G.GGGE <.GGG=

-. +7.T &0M0T(>

-. M G.GGG@
6+&M,(?)U-. M <.@D? U .GGG@ M G.GG%<ED
&+&M,(=)U-. M G.GGG U .GGG@ M G.GGG
K-. +7.T &0M0T(>

K,-&-. M <.GGGG

6+& M K,-&-. N .(<)U-. M <.GGGN%.G<=U.GGG@ M <.GGG@%%@
&+& M K,-&-. ; .(<)U-. M <.GGG;%.G<=U.GGG@ M %.FFF?EE@
8 + Chart#
XBA8 + Chart#
A%
D.F Attri.ute Charts in -enera$H p Charts in articu$ar
.ttribute control charts arise #hen items are compared #ith some standard and then are
classified as to #hether they meet the standard or not. The control chart is used to
determine if the rate of nonconforming product is stable and detect #hen a deviation from
stability has occurred. The argument can be made that a &+& should not exist, since rates
of nonconforming product outside the &+& is in fact a good thingL #e :.1T lo# rates of
nonconforming product. 7o#ever, if #e treat these &+& violations as simply another
search for an assignable cause, #e may learn for the drop in nonconformities rate and be
able to permanently improve the process.
!p" charts can be used #hen the subgroups are not of e*ual size. The np chart is used in
the more limited case of e*ual subgroups.
D.F.1 Steps in Constructing a p Chart
%. ,etermine the size of the subgroups needed. The size, n(i), has to be sufficiently
large to have defects present in the subgroup most of the time. 0f #e have some
idea as to #hat the historical rate of nonconformance, p, is #e can use the
follo#ing formula to estimate the subgroup size>
nM=4p
<. ,etermine the rate of nonconformities in each subgroup by using>
phat(i)Mx(i)4n(i)
#here>
phat(i)Mthe rate of nonconformities in subgroup i
x(i)Mthe number of nonconformities in subgroup i
n(i)M the size of subgroup i
A<
=. $ind pbarL there are k subgroups.
?. )stimate sigma;p if needed and determine the 6+& and &+&>
@. Plot the centerline, pbar, the &+& and 6+&, and the process measurements, the
phatZs.
A. 0nterpret the data to determine if the process is in control.
)*!mple:
1umber
,ay e8ects Tested Proportion
% %? <EA G.G?FG
< << <E% G.GDE=
= F =%G G.G<FG
? %F =%= G.GAGD
@ <% <F= G.GD%D
A %E =G@ G.G@FG
D %A =<< G.G?FD
E %A =%A G.G@GA
F <% <F= G.GD%D
%G %? <ED G.G?EE
%% %@ =GD G.G?EF
%< %A =<E G.G?EE
%= <% <FA G.GDGF
%? F <FA G.G=G?
%@ <@ =%D G.GDEF
%A %@ <FD G.G@G@
%D %? <E= G.G?F@
%E %= =<% G.G?G@
%F %G =%D G.G=%@
<G <% =GD G.GAE?
<% %F =%D G.G@FF
A=
<< <= =<= G.GD%<
<= %@ =G? G.G?F=
<? %< =G? G.G=F@
<@ %F =<? G.G@EA
<A %D <EF G.G@EE
<D %@ <FF G.G@G<
<E %= =%E G.G?GF
<F %F =%= G.GAGD
=G %< <EF G.G?%@
+alculations>
P-. M G.G@=F
6+& M pbar N =Us*rt(pbarU(%;pbar)4n(i))
&+& M pbar ; =Us*rt(pbarU(%;pbar)4n(i))
,ay +& 6+& &+& Proportion
% G.G@=F G.GF=EF<G?F G.G%=EGEAA% G.G?FG
< G.G@=F G.GF?<?AD<% G.G%=?@=FEF G.GDE=
= G.G@=F G.GF<=%GE<D G.G%@=EFEE= G.G<FG
? G.G@=F G.GF<%<AGAE G.G%@@D?A?< G.GAGD
@ G.G@=F G.GF=?%GE?= G.G%?<EFEAD G.GD%D
A G.G@=F G.GF<A<?DF@ G.G%@GD@F%@ G.G@FG
D G.G@=F G.GF%@ED=AE G.G%A%%==?< G.G?FD
E G.G@=F G.GF%F?=F?A G.G%@D@ADA? G.G@GA
F G.G@=F G.GF=?%GE?= G.G%?<EFEAD G.GD%D
%G G.G@=F G.GF=E<<<<F G.G%=EDE?E% G.G?EE
%% G.G@=F G.GF<?FE<EE G.G%@<G<?<< G.G?EF
%< G.G@=F G.GF%<?GA%F G.G%A?AGGF% G.G?EE
%= G.G@=F G.GF=<GFE@D G.G%??FGE@= G.GDGF
%? G.G@=F G.GF=<GFE@D G.G%??FGE@= G.G=G?
%@ G.G@=F G.GF%EE=E%? G.G%@E%AEFA G.GDEF
%A G.G@=F G.GF=%?=@? G.G%?@@D%D G.G@G@
%D G.G@=F G.GF?%G=D<? G.G%=@FAFEA G.G?F@
%E G.G@=F G.GF%A?A%G= G.G%AG@?AGD G.G?G@
%F G.G@=F G.GF%EE=E%? G.G%@E%AEFA G.G=%@
<G G.G@=F G.GF<?FE<EE G.G%@<G<?<< G.GAE?
<% G.G@=F G.GF%EE=E%? G.G%@E%AEFA G.G@FF
<< G.G@=F G.GF%@<EFGA G.G%A%D%EG? G.GD%<
<= G.G@=F G.GF<AEE@%D G.G%@G%<%F= G.G?F=
<? G.G@=F G.GF<AEE@%D G.G%@G%<%F= G.G=F@
<@ G.G@=F G.GF%?DGD%@ G.G%A<<FFF@ G.G@EA
<A G.G@=F G.GF=AE=ADE G.G%?G%DG=< G.G@EE
A?
<D G.G@=F G.GF=G%%FG? G.G%?AEEEGA G.G@G<
<E G.G@=F G.GF%E<=FAA G.G%@EDAD?? G.G?GF
<F G.G@=F G.GF<%<AGAE G.G%@@D?A?< G.GAGD
=G G.G@=F G.GF=AE=ADE G.G%?G%DG=< G.G?%@
p ; +hart>
D.C Attri.ute Charts in -enera$H c Charts in articu$ar
.ttribute control charts arise #hen items are compared #ith some standard and then are
classified as to #hether they meet the standard or not. The control chart is used to
determine if the rate of nonconforming product is stable and detect #hen a deviation from
stability has occurred. The argument can be made that a &+& should not exist, since rates
of nonconforming product outside the &+& is in fact a good thingL #e :.1T lo# rates of
nonconforming product. 7o#ever, if #e treat these &+& violations as simply another
search for an assignable cause, #e may learn for the drop in nonconformities rate and be
able to permanently improve the process.
The c +hart measures the number of nonconformities per ]unit] and is denoted by c. This
]unit] is commonly referred to as an inspection unit and may be ]per day] or ]per s*uare
foot] of some other predetermined sensible rate.
D.C.1 Steps in Constructing a c Chart
%. ,etermine cbar.
There are k inspection units and c(i) is the number of nonconformities in the ith
sample.
A@
<. (ince the mean and variance of the underlying Poisson distribution are e*ual,
Thus,
and the 6+& and &+& are>
=. Plot the centerline cbar, the &+& and 6+&, and the process measurements c(i).
?. 0nterpret the control chart.
E/amp$e#
1on;conforming
,ay )rrors4%GGG lines
% A
< D
= D
? A
@ E
A A
D @
E E
F %
%G A
%% <
%< @
%= @
%? ?
%@ =
%A =
%D <
%E G
%F G
<G %
<% <
<< @
<= %
<? D
<@ D
AA
<A %
<D @
<E @
<F E
=G E
+alculations>
+-. M ?.?AAD
6+& M cbar N =Us*rt(cbar) M %G.EGDG%=AA
&+& M cbar ; =Us*rt(cbar) M ;%.ED=AEG=<D M G
(#hen &+& R G, set &+& M G)
,ay +& 6+& &+& 1on+onforming
% ?.?AAD %G.EGDG%=AA G A
< ?.?AAD %G.EGDG%=AA G D
= ?.?AAD %G.EGDG%=AA G D
? ?.?AAD %G.EGDG%=AA G A
@ ?.?AAD %G.EGDG%=AA G E
A ?.?AAD %G.EGDG%=AA G A
D ?.?AAD %G.EGDG%=AA G @
E ?.?AAD %G.EGDG%=AA G E
F ?.?AAD %G.EGDG%=AA G %
%G ?.?AAD %G.EGDG%=AA G A
%% ?.?AAD %G.EGDG%=AA G <
%< ?.?AAD %G.EGDG%=AA G @
%= ?.?AAD %G.EGDG%=AA G @
%? ?.?AAD %G.EGDG%=AA G ?
%@ ?.?AAD %G.EGDG%=AA G =
%A ?.?AAD %G.EGDG%=AA G =
%D ?.?AAD %G.EGDG%=AA G <
%E ?.?AAD %G.EGDG%=AA G G
%F ?.?AAD %G.EGDG%=AA G G
<G ?.?AAD %G.EGDG%=AA G %
<% ?.?AAD %G.EGDG%=AA G <
<< ?.?AAD %G.EGDG%=AA G @
<= ?.?AAD %G.EGDG%=AA G %
<? ?.?AAD %G.EGDG%=AA G D
<@ ?.?AAD %G.EGDG%=AA G D
<A ?.?AAD %G.EGDG%=AA G %
<D ?.?AAD %G.EGDG%=AA G @
<E ?.?AAD %G.EGDG%=AA G @
<F ?.?AAD %G.EGDG%=AA G E
=G ?.?AAD %G.EGDG%=AA G E
AD
c + Chart#
D.D Attri.ute Charts in -enera$H np Charts in articu$ar
.ttribute control charts arise #hen items are compared #ith some standard and then are
classified as to #hether they meet the standard or not. The control chart is used to
determine if the rate of nonconforming product is stable and detect #hen a deviation from
stability has occurred. The argument can be made that a &+& should not exist, since rates
of nonconforming product outside the &+& is in fact a good thingL #e :.1T lo# rates of
nonconforming product. 7o#ever, if #e treat these &+& violations as simply another
search for an assignable cause, #e may learn for the drop in nonconformities rate and be
able to permanently improve the process.
The np +hart can be used for the special case #hen the subgroups are of e*ual size. Then
it is not necessary to convert nonconforming counts into the proportions phat(i). ather,
one can directly plot the counts x(i) versus the subgroup number i.
D.D.1 Steps in Constructing an np Chart
%. ,etermine the size of the subgroups needed. The size, n, has to be sufficiently
large to have defects present in the subgroup most of the time. 0f #e have some
idea as to #hat the historical rate of nonconformance, p, is #e can use the
follo#ing formula to estimate the subgroup size>
nM=4p
<. $ind find pbar.
AE
=. $ind the 6+& and &+& #here>
?. Plot the centerline pbar, the &+& and 6+&, and the process nonconforming
counts, the x(i)Zs.
@. 0nterpret the control chart. 2nly if a point is outside the N4; = sigma range is the
process considered to be out of control.
E/amp$e#
(ample
,ay 1on;conforming (ize
% %G %GG
< %< %GG
= %G %GG
? %% %GG
@ A %GG
A D %GG
D %< %GG
E %G %GG
F A %GG
%G %% %GG
%% F %GG
%< %? %GG
%= %A %GG
%? <% %GG
%@ <G %GG
%A %< %GG
%D %% %GG
%E A %GG
%F %G %GG
<G %G %GG
<% %% %GG
<< %% %GG
<= %% %GG
<? A %GG
<@ F %GG
+alculations>
P-. M G.%GEE
+& M %G.EEGG
6+& M nUpbar N =Us*rt(nUpbarU(%;pbar))
AF
&+& M nUpbar N =Us*rt(nUpbarU(%;pbar))
,ay +& 6+& &+& 1on+onforming
% %G.EEGG <G.<<%A?=@? %.@=E=@A?A< %G.GGGG
< %G.EEGG <G.<<%A?=@? %.@=E=@A?A< %<.GGGG
= %G.EEGG <G.<<%A?=@? %.@=E=@A?A< %G.GGGG
? %G.EEGG <G.<<%A?=@? %.@=E=@A?A< %%.GGGG
@ %G.EEGG <G.<<%A?=@? %.@=E=@A?A< A.GGGG
A %G.EEGG <G.<<%A?=@? %.@=E=@A?A< D.GGGG
D %G.EEGG <G.<<%A?=@? %.@=E=@A?A< %<.GGGG
E %G.EEGG <G.<<%A?=@? %.@=E=@A?A< %G.GGGG
F %G.EEGG <G.<<%A?=@? %.@=E=@A?A< A.GGGG
%G %G.EEGG <G.<<%A?=@? %.@=E=@A?A< %%.GGGG
%% %G.EEGG <G.<<%A?=@? %.@=E=@A?A< F.GGGG
%< %G.EEGG <G.<<%A?=@? %.@=E=@A?A< %?.GGGG
%= %G.EEGG <G.<<%A?=@? %.@=E=@A?A< %A.GGGG
%? %G.EEGG <G.<<%A?=@? %.@=E=@A?A< <%.GGGG
%@ %G.EEGG <G.<<%A?=@? %.@=E=@A?A< <G.GGGG
%A %G.EEGG <G.<<%A?=@? %.@=E=@A?A< %<.GGGG
%D %G.EEGG <G.<<%A?=@? %.@=E=@A?A< %%.GGGG
%E %G.EEGG <G.<<%A?=@? %.@=E=@A?A< A.GGGG
%F %G.EEGG <G.<<%A?=@? %.@=E=@A?A< %G.GGGG
<G %G.EEGG <G.<<%A?=@? %.@=E=@A?A< %G.GGGG
<% %G.EEGG <G.<<%A?=@? %.@=E=@A?A< %%.GGGG
<< %G.EEGG <G.<<%A?=@? %.@=E=@A?A< %%.GGGG
<= %G.EEGG <G.<<%A?=@? %.@=E=@A?A< %%.GGGG
<? %G.EEGG <G.<<%A?=@? %.@=E=@A?A< A.GGGG
<@ %G.EEGG <G.<<%A?=@? %.@=E=@A?A< F.GGGG
np + Chart#
Buestions>
The follo#ing table gives the number of missing rivets noted in ne#ly fabricated buses.
DG
-us
1o.
% < = ? @ A D E F %G %% %< %= %? %@ %A %D %E %F <G
1o.of
missing
rivets
A D D A E A @ E % A < @ @ ? = =< < G G %
<% << <= <? <@ <A <D <E <F =G
< @ % D D % @ @ E E
+onstruct c;chart and comment on the process of fabrication
Chapter 1B
Tota$ ?ua$it& management
1B.1 Introduction
Total Buality Management (TBM), a buzz#ord phrase of the %FEGZs, has been killed and
resurrected on a number of occasions. The concept and principles, though simple seem to
be creeping back into existence by ]bits and pieces] through the evolution of the
0(2FGG% Management Buality (ystem standard.
+ompanies #ho have implemented TBM include $ord Motor +ompany, Phillips
(emiconductor, ('& +arbon, Motorola and Toyota Motor +ompany.
The latest changes coming up for the 0(2 FGG%><GGG standard"s ]Process Model] seem to
complete the embodiment. TBM is the concept that *uality can be managed and that it is
a process. The follo#ing information is provided to give an understanding of the key
elements of this process.
Tota$ >ua$it& Management 1T>M4
Tota$ M Buality involves everyone and all activities in the company.
>ua$it& M +onformance to e*uirements (Meeting +ustomer e*uirements).
Management M Buality can and must be managed.
T>M M . process for managing *ualityL it must be a continuous #ay of lifeL a philosophy
of perpetual improvement in everything #e do.
D%
T>M Compared to ISO DBB1
0(2 FGGG is a Buality (ystem Management (tandard. TBM is a philosophy of perpetual
improvement. The 0(2 Buality (tandard sets in place a system to deploy policy and
verifiable ob8ectives. .n 0(2 implementation is a basis for a Total Buality Management
implementation. :here there is an 0(2 system, about D@ percent of the steps are in place
for TBM. The re*uirements for TBM can be considered 0(2 plus. .nother aspect
relating to the 0(2 (tandard is that the proposed changes for the next revision (%FFF) #ill
contain customer satisfaction and measurement re*uirements. 0n short, implementing
TBM is being proactive concerning *uality rather than reactive.
T>M as a Foundation
T>M is the foundation for activities #hich includeL
Meeting +ustomer e*uirements
educing ,evelopment +ycle Times
9ust 0n Time4,emand $lo# Manufacturing
0mprovement Teams
educing Product and (ervice +osts
0mproving .dministrative (ystems Training
The Ten Steps to T>M are as )o$$o,s#
%. Pursue 1e# (trategic Thinking
<. /no# your +ustomers
=. (et True +ustomer e*uirements
?. +oncentrate on Prevention, 1ot +orrection
@. educe +hronic :aste
A. Pursue a +ontinuous 0mprovement (trategy
D. 6se (tructured Methodology for Process 0mprovement
E. educe 5ariation
F. 6se a -alanced .pproach
%G. .pply to .ll $unctions
rincip$es o) T>M
The rincip$es o) T>M are as )o$$o,s#
%. Buality can and must be managed.
<. )veryone has a customer and is a supplier.
=. Processes, not people are the problem.
?. )very employee is responsible for *uality.
@. Problems must be prevented, not 8ust fixed.
A. Buality must be measured.
D. Buality improvements must be continuous.
D<
E. The *uality standard is defect free.
F. 'oals are based on re*uirements, not negotiated.
%G. &ife cycle costs, not front end costs.
%%. Management must be involved and lead.
%<. Plan and organize for *uality improvement.
rocesses must .e Managed and Impro*ed
Processes must be managed and improved_ This involves>
,efining the process
Measuring process performance (metrics)
evie#ing process performance
0dentifying process shortcomings
.nalyzing process problems
Making a process change
Measuring the effects of the process change
+ommunicating both #ays bet#een supervisor and user
Le& to >ua$it&
The key to improving *uality is to improve processes that define, produce and support
our products.
.ll people #ork in processes.
People
'et processes ]in control]
:ork #ith other employees and managers to identify process problems and
eliminate them
M!n!+ers !nd,or S&per#isors -or.
on Processes Provide training and tool resources
Measure and revie# process performance (metrics)
0mprove process performance #ith the help of those #ho use the process
D=

T>M rocess Impro*ement and ro.$em So$*ing Se?uence
LAN
(P&.1 . +7.1'))
"O
(0MP&)M)1T
T7)
+7.1'))
CIECL
(2-()5) T7)
)$$)+T()
ACTION
()M-), T7)
$0K 01T2
T7) P2+)((
$2 '22,)
"EFINE
TIE
8OBLEM
I"ENTIFR
OSSIBLE
CAGSES
E3ALGATE
OSSIBLE
CAGSES
MALE
A
CIAN-E
TEST
TIE
CIAN-E
TALE
E8MANENT
ACTION
%. ecognize that
#hat you are
doing is a
]P2+)((]
A.
]-.01(T2M]
#hat is causing
the problem.
E. ,etermine
the
relationship
bet#een
cause and
%G. ,etermine
#hat change
#ould help
Oour
kno#le
%%. ,etermine
#hat change
#orked
(confirmation).
7istograms
%<. )nsure the
fix is embedded
in the process
and that the
resulting
D?
<. 0dentify the
commodity
being processed.
; Process
0nference
=. ,efine some
measurable
characteristics of
value to the
commodity.
?. ,escribe the
]P2+)((]
o Process $lo#
.nalysisZs
o $lo# charts
o &ist of steps
@. 0dentify the
]-ig] problem
o -rainstorming
o +hecklists
o Pareto analysis
D. ,etermine #hat
past data sho#s.
o $re*uency
distribution
o Pareto charts
o +ontrol charts
+ sampling
effect
o (catter
diagrams
o egression
analysis
F. ,etermine
#hat the
process is
doing no#
o +ontrol
charts
; sampling
dge
of the
process
(catter
diagram
s
+ontrol
+harts
+
samplin
g
Pareto
analysis
UUUUThen
make
the change.
+ontrol
charts
+ sampling
(catter
diagrams
process is used.
+ontinue to
monitor the
process to
ensure>
.. The problem
is fixed for
good.
and
-. The process
is good enough
o +ontrol charts
+ sampling
UUUUTo ensure
continuous
improvement,
return
to step @.
D@
ISO DBB1 Features#
The 0(2 FGGG family of international *uality management standards and guidelines has
earned a global reputation as the basis for establishing *uality management systems.
The familiar three standards 0(2 FGG%, 0(2 FGG< and 0(2 FGG= have been integrated into
the ne# 0(2 FGG%><GGG. 0(2 FGG%><GGG specifies re*uirements for a *uality management
system for any organization that needs to demonstrate its ability to consistently provide
product that meets customer and applicable regulatory re*uirements and aims to enhance
customer satisfaction. 0(2 FGG%><GGG has been organized in a user;friendly format #ith
terms that are easily recognized by all business sectors. The standard is used for
certification4registration and contractual purposes by organizations seeking recognition of
their *uality management system.
The greatest value is obtained #hen you use the entire family of standards in an
integrated manner. 0t is suggested that, beginning #ith 0(2 FGGG><GGG, you adopt 0(2
FGG%><GGG to achieve a first level of performance. The practices described in 0(2
FGG?><GGG may then be implemented to make your *uality management system
increasingly effective in achieving your o#n business goals. 0(2 FGG%><GGG and 0(2
FGG?><GGG have been formatted as a consistent pair of standards to facilitate their use.
6sing the standards in this #ay #ill also enable you to relate them to other management
systems (e.g. environmental), many sector specific re*uirements (such as 0(24T(4%AF?F
in the automotive industry) and #ill assist in gaining recognition through national a#ard
programmes.
0(2 FGG%><GGG is used if you are seeking to establish a management system that
provides confidence in the conformance of your product to established or specified
re*uirements. 0t is no# the only standard in the 0(2 FGGG family against #hose
re*uirements your *uality system can be certified by an external agency. The standard
recognizes that the #ord ]product] applies to services, processed material, hard#are
and soft#are intended for, or re*uired by, your customer.
There are five sections in the standard that specify activities that need to be
considered #hen you implement your system. Oou #ill describe the activities you use
to supply your products and may exclude the parts of the Product ealization section
that are not applicable to your operations. The re*uirements in the other four
sections C Buality management system, Management responsibility, esource
management and Measurement, analysis and improvementC apply to all organizations
and you #ill demonstrate ho# you apply them to your organization in your *uality
manual or other documentation.
Together, the five sections of 0(2 FGG%><GGG define #hat you should do consistently
DA
to provide products that meets customer and applicable statutory or regulatory
re*uirements. 0n addition, you #ill seek to enhance customer satisfaction by
improving your *uality management system.
0(2 FGG?><GGG is used to extend the benefits obtained from 0(2 FGG%><GGG to all
parties that are interested in or affected by your business operations. 0nterested parties
include your employees, o#ners, suppliers and society in general.
0(2 FGG%><GGG and 0(2 FGG?><GGG are harmonized in structure and terminology to
assist you to move smoothly from one to the other. -oth standards apply a process
approach. Processes are recognized as consisting of one or more linked activities that
re*uire resources and must be managed to achieve predetermined output. The output
of one process may directly form the input to the next process and the final product is
often the result of a net#ork or system of processes. The eight Buality Management
Principles stated in 0(2 FGGG><GGG and 0(2 FGG?><GGG provide the basis for the
performance improvement outlined in 0(2 FGG?><GGG.
The nature of your business and the specific demands you have #ill determine ho#
you apply the standards to achieve your ob8ectives.
Iere are a )e, e/amp$es#
E/amp$e 1
. metal parts fabricating company used 0(2 FGGG><GGG to develop a plan to
implement their *uality management system. :hen they #ere ready, they prepared a
*uality manual and *uality system procedures as re*uired by 0(2 FGG%><GGG,
excluding the re*uirements covering product design and development because their
products are made to designs prepared by their customers. &ater, in order to bid on
the supply of parts to a ma8or automotive company, they adapted their *uality system
to meet the sector specific re*uirements of 0(24T( %AF?F.
E/amp$e !
. #elfare agency decided to establish a *uality improvement strategy. 0t adopted 0(2
FGG?><GGG as the basis for planning and implementing its system. The agency found
that 0(2 FGGG><GGG provided very useful additional guidance and plans to seek
certification to 0(2 FGG%><GGG to gain more credibility.
E/amp$e (
. #ashing machine manufacturer had a #ell;established company culture of
continual improvement and effective production control. The management decided to
improve the companyZs development processes and to implement 0(2 FGG%><GGG to
obtain certification for commercial purposes. The company used 0(2 FGG?><GGG to
DD
guide its improvement processes and 0(2 %GGGA>%FFD to develop a pro8ect
management plan.
E/amp$e 9
. large chemical processing company #as re*uired by its ma8or customers to gain
registration4certification to 0(2 FGG%><GGG. 0n order to obtain additional benefits,
company leadership planned a comprehensive management strategy based on 0(2
FGGG><GGG and 0(2 FGG?><GGG. . thorough revie# of their business processes
indicated that all elements of 0(2 FGG%><GGG #ere applicable to their *uality
management system. The company used 0(2 %GG%=>%FFF to guide the development
of *uality documentation in its various production divisions and 0(2 %GG%@>%FFF for
guidance in the preparation of training plans for their employees.
E/amp$e :
. firm of international la#yers #anted to improve their client management processes
and to achieve registration4certification to 0(2 FGG%><GGG. Their *uality management
system provides for the design and development of ne# services such as international
tax planning and modifying traditional services to meet the re*uirements of ne# or
amended legislation. They included purchasing control to cover the selection of
computer hard#are and soft#are, as #ell as purchasing the services of specialist
la#yers as needed. .fter successfully implementing 0(2 FGG%><GGG, they used the
self;assessment guidelines of 0(2 FGG?><GGG to monitor their progress as they
improved their *uality management system.
E/amp$e =
. computer soft#are developer serving a niche market recognized that as their user
base expanded they #ould be faced #ith issues concerning product management and
configuration control. +hanges to base products, user hard#are and regulatory
re*uirements #ere compounding customer service issues. 0(2 FGG?><GGG provided
the guidance they needed to establish documented procedures to control process
change and improvement. 0(2 %GGGA>%FFD and 0(2 %GGGD>%FF@ provided additional
assistance as they managed the pro8ect and prepared procedures for configuration
management. They later ac*uired another soft#are developer and #ere able to use
their *uality management system to integrate the ac*uisition into their o#n structure
very *uickly #ith a minimum of disruption to customers.
E/amp$e F
. bank decided to implement a *uality management system for its on;line 0nternet
banking services. They ensured that their *uality manual made clear that their other
conventional banking services #ere not included in their *uality management system.
:hile adopting the re*uirements of 0(2 FGG%><GGG, the bank obtained guidance from
DE
0(2 FGGG><GGG to interpret #ords and phrases used in the standard for their
application. They applied all the re*uirements of +lause D, recognizing that design
and development is an important part of creating ne# service processes. The bank
used 0(2 %GG%=>%FF@ to prepare their documentation, #hich they posted on their
internal computer net#ork to ensure current procedures are available to their staff.
Imp$ementing &our ISO DBB1#!BBB ?ua$it& management s&stem
1. Identi)& the goa$s &ou ,ant to achie*e
Typical goals may be>
-e more efficient and profitable
Produce products and services that consistently meet customer re*uirements
.chieve customer satisfaction
0ncrease market share
Maintain market share
0mprove communications and morale in the organization
educe costs and liabilities
0ncrease confidence in the production system
!. Identi)& ,hat others e/pect o) &ou
These are the expectations of interested parties (stakeholders) such as>
+ustomers and end users
)mployees
(uppliers
(hareholders
(ociety
(. O.tain in)ormation a.out the ISO DBBB )ami$&
$or general information, look to this brochure
$or more detailed information, see 0(2 FGGG><GGG, 0(2 FGG%><GGG and 0(2
FGG?><GGG
$or supporting information, refer to the 0(2 :eb site
$or implementation case studies and ne#s of 0(2 FGGG developments #orld#ide,
read the 0(2 publication )*O %anagement *ystems.
9. App$& the ISO DBBB )ami$& o) standards in &our management s&stem.
DF
,ecide if you are seeking certification that your *uality management system is in
conformance #ith 0(2 FGG%><GGG or if you are preparing to apply for a national *uality
a#ard.
6se 0(2 FGG%><GGG as the basis for certification
6se 0(2 FGG?><GGG in con8unction #ith your national *uality a#ard criteria to
prepare for a national *uality a#ard
:. O.tain guidance on speci)ic topics ,ithin the ?ua$it& management s&stem
These topic;specific standards are>
0(2 %GGGA for pro8ect management
0(2 %GGGD for configuration management
0(2 %GG%< for measurement systems
0(2 %GG%= for *uality documentation
0(24T %GG%? for managing the economics of *uality
0(2 %GG%@ for training
0(24T( %AF?F for automotive suppliers
0(2 %FG%% for auditing
=. Esta.$ish &our current statusH determine the gaps .et,een &our ?ua$it&
management s&stem and the re?uirements o) ISO DBB1#!BBB
Oou may use one or more of the follo#ing>
(elf assessment
.ssessment by an external organization
F. "etermine the processes that are needed to supp$& products to &our customers
evie# the re*uirements of the 0(2 FGG%><GGG section on Product ealization to
determine ho# they apply or do not apply to your *uality management system includingL
+ustomer related processes
,esign and4or development
Purchasing
Production and service operations
+ontrol of measuring and monitoring devices
C. "e*e$op a p$an to c$ose the gaps in step = and to de*e$op the processes in step F
0dentify actions needed to close the gaps, allocate resources to perform these actions,
assign responsibilities and establish a schedule to complete the needed actions. 0(2
FGG%><GGG Paragraphs ?.% and D.% provide the information you #ill need to consider
#hen developing the plan.
EG
D. Carr& out &our p$an
Proceed to implement the identified actions and track progress to your schedule
1B. Gndergo periodic interna$ assessment
6se 0(2 %FG%% for guidance in auditing, auditor *ualification and managing audit
programmes
11. "o &ou need to demonstrate con)ormance'
)# yes+ go to step ,-
)# no+ go to step ,.
Oou may need or #ish to sho# conformance (certification4registration) for various
purposes, for example>
+ontractual re*uirements
Market reasons or customer preference
egulatory re*uirements
isk management
To set a clear goal for your internal *uality development (motivation)
1!. Gndergo independent audit
)ngage an accredited registration4certification body to perform an audit and certify that
your *uality management system complies #ith the re*uirements of 0(2 FGG%><GGG.
1(. Continue to impro*e &our .usiness
evie# the effectiveness and suitability of your *uality management system. 0(2
FGG?><GGG provides a methodology for improvement
TTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTTT
8EFE8ENCES FO8 TIIS COG8SE BOOL#
%. (upply +hain Management; (unil +hopra and Peter Meindl, Pearson )ducation
<. Production and 2perations Management; .Paneerselvam, P70
=. Production and 2perations Management; (.1.+hary, Tata;Mc'ra# 7ill
?. 2perations Management for +ompetitive .dvantage;+hase, .*uilano and 9acobs,
Tata;Mc'ra# 7ill
@. ,esigning and Managing the (upply +hain, ,avid (imchi;&evi, Philip /aminsky
and (imchi;&evi, Tata;Mc'ra# 7ill
E%
>ueuing mode$s in capacit& p$anning )or ser*ice industries
Bueuing theory is often useful for designing service systems. :aiting lines have a
tendency to form in a #ide variety of service systems (.irport ticket counter, telephone
calls to a cable television company, 7ospitals, ail#ay reservation counter). The lines are
symptoms of bottleneck operations. Bueuing theory is useful in helping managers choose
a capacity level that #ill be cost effective through balancing the cost of having customers
#ait #ith the cost of providing additional capacity. 0t can aid in the determination of
expected costs for various levels of service capacity.
1o doubt those #aiting in line #ould all agree that the solution to the problem is obvious>
simply add more servers or else do something to speed up service. :hile both ideas may
be potential solutions, there are certain subtleties that must be dealt #ith. $or on thing,
most service systems have the capacity to process more customers over the long run than
they are called upon to process. 7ence, the problem of #aiting is a short;term
phenomenon. The other side of the coin is that at certain times the system is empty, and
servers are idle #aiting for customers. Thus by increasing the service capacity, the server
idle time #ould increase. +onse*uently, in designing service systems, the designer must
#eigh the cost of providing a given level of service capacity against the potential cost of
having customers #ait for service.
The planning and analysis of service capacity fre*uently lends itself to *ueuing theory,
#hich is a mathematical approach to the analysis of #aiting line problem. These #aiting
lines are commonly found #herever customers arrive randomly for services. (e.g.
supermarket checkouts, theaters, toll gates, post;offices, ation shops, ). 0n business
situations, the customers are not people, but orders #aiting to be filled, trucks #aiting to
be unloaded, 8obs #aiting to be processed, or e*uipment #aiting for repairs.
:hy is there #aitingC
Many people are surprised #hen they learn that #aiting lines tend to form even though a
system is basically under loaded. $or example, a fast;food restaurant may have the
capacity to handle an average of <GG orders per hour and yet experience #aiting line even
though the average number of orders is only %@G per hour. The key #ord is average. 0n
reality, customers arrive at random intervals rather than at evenly spaced intervals, and
some orders take longer to fill than others. 0n other #ords, both arrivals and service times
exhibit a high degree of variability. .s a result of this variability there are times #hen the
system becomes temporarily overloaded, and that gives rise to #aiting lines. -y the same
token, there are other times #hen the system is idle because there are no customers.
Thus, although a system may be under loaded from a macro standpoint. -ecause of
variability in arrivals and service there #ill be times #hen the system is over loaded
from a micro standpoint.
E<
'oal of Bueuing analysis>
The goal of *ueuing is essentially to minimize total costs. There are t#o basic categories
of cost in a *ueuing situation> those associated #ith customers #aiting for service and
those associated #ith capacity. +apacity costs relate to maintaining the ability to provide
to provide capacity. (the number of check;outs in a super market, number of reservation
counters) :aiting costs include the salaries paid to employees, #aiting space cost and
customer dissatisfaction.
The goal of *ueuing analysis is to balance the cost of providing a level of service capacity
and the cost of customers #aiting for service.
+ustomer source
:aiting line (ervice facility
E=
+
2
(
T
T+
+ost of customer
#aiting
+ost of capacity
(ervice capacity
S&stem characteristics>
There are numerous *ueuing models from #hich an analyst can choose. Model choice is
dependent on the characteristics of the system under investigation. The main
characteristics are>
%. The population source
<. The number of servers (+hannels)
=. .rrival and service time patterns
?. The *ueue discipline (order of service)
Population source>
There are t#o possibilities> infinite source and finite source populations. 0n an infinite
source situation, the potential number of customers greatly exceeds system capacity.
((ervice is unrestricted;supermarkets, restaurants, tollgates), :hen the potential clients is
limited, a finite source situation exists. (. repair person for a group of machines).
1umber of servers>
The capacity of *ueuing system is a function of the capacity of each server and the
number of servers being used. (ystems can be either single or multiple channels.
.rrival and (ervice patterns>
Most commonly used models assume that the customer arrival rate can be described by a
Poisson distribution and that of the service time can be described by a negative
exponential distribution.
Bueue discipline>
.ll of the models assume that service is provided on a firs;come, first; served basis.
(-anks, stores, .TMs,) )xamples of systems that do not follo# by $+$( rule include
hospital emergency rooms, rush orders in factory and computer processing of programs.
B6)601' M2,)&(
%. 0nfinite source, single channel, exponential service time
<. 0nfinite source, single channel, constant service time
=. 0nfinite source, multiple channel, exponential service time
&ike above another set of three models are there for finite source.
(ymbols used in infinite source models>
E?
; customer arrival rate
; service time
&*; the average number of customers #aiting for service
&s; the average number of customers in the system
;The system utilization
:* P the average time customers #ait in the line.
:s P the average time customers spend in the system
%4 ; service time
M; the number of servers4channels
&max P the maximum expected number #aiting in line
Measures o) s&stem per)ormance#
1.
The a*erage num.er .eing ser*ed 6r7 0 5
!.
The a*erage num.er in the s&stem Ls 0 L? S r
(.
The a*erage time in $ine %? 0 L?5r
9.
The a*erage time in the s&stemH inc$uding ser*ice time# %s0 %? S15
:.
S&stem uti$i;ation# 0 51M4
Formu$as )or Basic sing$e ser*er mode$#
=.
A*erage num.er in $ine# L? M
!
51 + 4
F.
ro.a.i$it& o) ;ero units in the s&stem B 0 1+1 54
C.
ro.a.i$it& o) 6n7 units in the s&stem# n 0 B1 54
n
E/amp$e#
.n airline is planning to open a satellite ticket desk in a ne# shopping plaza, staffed by
one ticket agent. 0t is estimated that re*uests for tickets and information #ill average %@
per hour and re*uests #ill have a Poisson distribution. (ervice time is assumed as
exponentially distributed. Previous experience #ith similar satellite operations suggests
that mean service time should average about three minutes per re*uest. ,etermine the
follo#ing.
%. (ystem utilization
<. Percent of time the server (agent ) #ill be idle
=. The expected number of customers #aiting to be served.
?. The average time customers #ill spend in the system
@. The probability of zero customers in the system and the probability of four
customers in the system
So$ution#
0 1: per hour
0 15ser*ice time 011 customer5( minutes4 /=B minutes per hour 0 !B customers per
hour
E@
S&stem uti$i;ation# 0 51M40 1:511!B40B.F:0F:T
Id$e time0 1+ 0 !:T
L? M
!
51 + 4 0 1:/1:5!B1!B+1:40!.!: customers
%s0 %? S150 L?5 S 15 0!.!:51: S15!B 0B.!B hours or 1! minutes
B 0 1+1 540 1+1:5!B 0 B.!:
90 B.!: 11:5!B4
9
0B.FD
12T)>
The other models can be learned by referring any 2 books and operations management
books.
KGST IN TIME SRSTEMS
The 8ust;in;time approach attempts to minimize inventories. 0t accomplishes this by
ordering and producing very small batches. 0deally, as each order of ra# material arrives
from a vendor, it goes directly into production, and as each batch is finished at one #ork
center, it is moved to the follo#ing #ork center #here the previous order has 8ust been
completed and #ork is begun on the ne# order. 7ence, inventories of ra# materials and
purchased parts are purposely kept as small as possible, as are inventories of partially
completed goods bet#een operations. 1aturally, this means that a disruption any#here in
the system can easily cause the entire system to come to a halt since there is little slack
bet#een operations. 7o#ever the 9apanese see this as an opportunity to learn #here the
#eak points are in the system and to correct those problems so they do not occur. -y #ay
of contrast, most 6.(. manufacturers carry safety stock W8ust in caseH..X .ccording to the
90T philosophy, buffers of inventory tend to cover up or hide problems that keep recurring
and are never resolved, partly because they are not al#ays obvious and partly because the
presence of inventory seems to make the problems less serious. 1eedless to say, the
additional inventory carries a burden in terms of holding costs.
$actors that enable 90T system>
%. 0n order to maintain a 90T system #ith little inventory and small run sizes, it is
necessary to have *uick, lo#;cost changeovers and setups.
<. 0n;process inventories are kept small by running small run sizes. The ability to
*uickly change over e*uipment and have relatively lo#;cost setups permits them
to do this and keep inventories lo# in the process.
=. The #illingness to commit the time, effort and money to correct problems at the
source is essential to the success of the system.
The 9apanese have had a national goal of full employment through industrialization since
:orld :ar 00. The strategy for obtaining market dominance has been targeted to specific
EA
product areas;in particular, automotive assembly and consumer electronics. They directed
their best engineering talent to the shop floor, not to product design activities. They also
#orked to improve product *uality and reliability above #hat competitors could supply.
+entral to this effort #ere t#o philosophies>
%. )limination of #aste
<. espect for people
ELIMINATION OF %ASTE
Toyota"s $u8io +ho identifies seven prominent types of #aste to be eliminated.
%. :aste from over production
<. :aste of #aiting time
=. Transportation #aste
?. 0nventory #aste
@. Processing #aste
A. :aste of motion
D. :aste from product defects.
This definition of 90T leaves no room for surplus or safety stock. 1o safety stocks are
allo#ed because if you cannot use it no#, you do not need to make it no#L that #ould be
#aste. 7idden inventory in storage areas, transit systems, carousels, and conveyors is a
key target for inventory reduction.
The seven elements that address elimination of #aste are,
%. $ocused factory
<. 'roup technology
=. Buality at source
?. 90T Production
@. 6niform plant loading
A. /anban production control system
D. Minimized setup time
FOCGSE" FACTO8R NET%O8LS
The 9apanese build small;specialized plants rather than large vertically integrated
manufacturing facilities. Plants designed for one purpose can be constructed and operated
more economically.
-8OG TECINOLO-R
This is a philosophy in #hich similar parts are grouped into families, and the processes
re*uired to make the parts are arranged in a specialized #ork cell. 0nstead of transferring
8obs from one department to another to specialized #orkers, the 9apanese consider all
operations re*uired to make a part and group those machines together. The 'T removes
movements and *ueue time bet#een operations, reduce inventory and reduce the number
ED
of employees re*uired. :orkers, ho#ever, must be flexible to run several machines and
processes. ,ue to their advanced skill level, these #orkers have increased 8ob security.
>GALITR AT TIE SOG8CE
0t means do it right the first time and, #hen something goes #rong stop the process or
assembly line immediately. $actory #orkers become their o#n inspectors, personally
responsible for the *uality of their output.
KIT 8O"GCTION
90T means producing #hat is needed and no more. .nything over the minimum amount
necessary is vie#ed as #aste, because effort and material expended for something not
needed no# cannot be utilized no#. This is in contrast to relying on extra material 8ust in
case something goes #rong.
90T is typically applied to repetitive manufacturing, #hich is #hen the same or similar
items are made one after another. 0t does not re*uire large volumes and ca n be applied to
any repetitive segments of a business regardless of #here they appear. 6nder 90T the ideal
lot size is one.
GNIFO8M LANT LOA"IN-
(moothing the production flo# to dampen the reaction #aves that normally occur in
response to schedule variations is called uniform plant loading. :hen a change is made in
a final assembly, the changes are magnified throughout the line and the supply chain. The
only #ay to eliminate the problem is to make ad8ustments as small as possible by setting
a firm monthly production plan for #hich the output rate is frozen.
LANBAN 8O"GCTION CONT8OL SRSTEM#
. kanban control system uses a signaling device to regulate 90T flo#s. /anban means
WsignX or instruction card in 9apanese. 0n a paperless control system, containers can be
used instead of cards. This kanban pull system tries co;ordinate bet#een departments or
#ork centers by means of signaling #hen the next lot should be produced.
CONCET OF GLL AN" GSI 8O"GCTION SRSTEM
The concept of pull in lean production means to respond to the pull, or demand, of the
customer. &ean manufacturers design their operations to respond to the ever;changing
re*uirements of customers. Those able to produce to the pull of customers do not need to
manufacture goods that traditional batch;and;*ueue manufacturers must rely on. The
planning for delivery of product to customers is less troublesome, and demand becomes
more stable if customers have confidence in kno#ing that they can get #hat they #ant
#hen they #ant it.
/anban is a 9apanese #ord that means ]instruction card]. /anbans are manual pull
devices that allo# an efficient means to transfer parts from one department to another and
automatically reorder products using minimum4maximum inventory levels. . /anban is a
EE
signal, such as an empty container returned to the start of the assembly line, that signals
the need for replenishment of m!teri!ls to ! &ser.
/anbans are used in ]pull] manufacturing systems, #here product is
manufactured to the pull of market;driven demand.
/anban systems must be convenient and easy to use
Pull systems react to needs, they donZt anticipate them
(uccessfully deployed /anbans deliver the right amount of material to the right
place exactly #hen it is needed.
(everal pull techni*ues can be used for different products at the same #orksite.
'reat speed can be achieved in manufacturing, and product is not manufactured
#hen a need does not exist.
. kanban or pull system means providing the #orkers #ith #hat they need #hen
they need it ; tools, soft#are, capital e*uipment, access, feedback, or the
opportunity to participate.
/anbans or cards are used #hen the move time and distance bet#een producing
and consuming departments are significant.
o )ach card controls a specific *uantity of parts
o +ards are returned to the producing department after parts are consumed
triggering production of the next batch
o (ingle card systems are used #hen the products are able to be resupplied
prior to running out
o Multiple card systems are used if the producing #orksite produces several
products or if the lot size is different from the move size
Physical /anbans, /anban s*uares, or shelf reserves are used #hen the producing
and consuming #orksites are physically ad8acent.
o Must have relatively close proximity
o 6sed #hen ma8or *ueue of parts is maintained in the producing
department. Maximum and minimum *ueue size can coordinate
production bet#een #orksites #ith different capacities
o Physical area holds enough space to cover variations in rates
o 'oal of the producing department is to fill the space reserved for the part.
KANAN vs! A"C#$AN%$&'E'E
The unpredictable and expensive batch;and;*ueue method of manufacturing, coupled
#ith unreliable forecasting associated #ith traditional production models, is replaced
#ith reliable, predictable, /anban systems.
/E0EF)T*:
.chieves shop floor control of inventory and production scheduling #ithout the
aid of materials re*uirements planning (MP)
&inks producing and consuming processes #hile often reducing :0P by @GY
0ncreases productivity
educes inventory and floor space
EF
,ramatically improves lead times
(ets up point;of;use material storage and better interfaces #ith planning systems
(ets lot sizes #ith capacity rather than economic order *uantity theory
()A* AN% P#+*)S)P#,
The Theory of +onstraints (or T2+ as it is called) is a relatively recent development in
the practical aspect of making organizational decisions in situations in #hich constraints
exist. ,r. )liyahu M. 'oldratt first described the theory in his novel, The 1oal( 0n many
organizations, T2+ and T2+ logic form ma8or portions of that organizationZs philosophy
of continuous improvement.
The Theory of +onstraints has been used at three different levels> Production
Management ; T2+ #as initially applied here to solve problems of bottlenecks,
scheduling, and inventory reduction.
Throughput .nalysis ; .pplication of T2+ has caused a shift from cost;based decision
making to decision;making based on continuous improvement of processes in #hich
system throughput, system constraints, and statistically determined protective capacities
at critical points are key elements.
Theory of +onstraints &ogical Processes ; This third level is the general application of
T2+ reasoning to attack a variety of process problems #ithin organizations. T2+ logic is
applied to identify #hat factors are limiting an organization from achieving its goals,
developing a solution to the problem, and getting the individuals in the process to invent
the re*uisite changes for themselves.
. constraint is anything in an organization that limits it from moving to#ard or achieving
its goal. 2f course, this assumes that an appropriate goal has been defined. $or most
business organizations the goal is to make money no# as #ell as in the future. There are
t#o basic types of constraints> physical constraints and non;physical constraints. .
physical constraint is something like the physical capacity of a machine. . non;physical
constraint might be something like demand for a product, a corporate procedure, or an
individualZs paradigm for looking at the #orld.
The steps in applying T2+ are as follo#s>
%. 0dentify the systemZs constraints. 2f necessity this included prioritization so that
8ust the ones that really limit system progress to#ard the goal.
<. ,ecide ho# to exploit the systemZs constraints. 2nce #e have decided ho# to
manage the constraints #ithin the system, ho# about the ma8ority of the resources
that are not constraintsC The ans#er is that #e manage them so that they 8ust
provide #hat is needed to match the output of the constrained resources. :e
FG
1)5) let them supply more output than is needed because doing so moves us
no closer to the goal.
=. (ubordinate everything else to the above decision in (tep <. (ince the constraints
are keeping us from moving to#ard our goal, #e apply all of the resources that #e
can to assist in breaking them. +onstraints are not acts of 'od. 0n practically all
cases their limiting impact can be reduced or eliminated.
?. )levate the systemZs constraints. 0f #e continue to #ork to#ard breaking a
constraint (also called elevating a constraint) at some point the constraint #ill no
longer be a constraint. The constraint #ill be broken.
@. 0f the constraint is broken, return to (tep %. :hen that happens, there #ill be
another constraint, some#here else in the system that is limiting progress to the
goal.
The process must be reapplied, perhaps many times. 0t is very important not to let inertia
become a constraint. Most constraints in organization are of their o#n making. They are
the entrenched rules, policies, and procedures that have developed over time. Many
times, #hen #e finally break a constraint, #e do not go back and revie# and change the
rules and policies that caused the constraint initially. Most constraints in organizations
today are policy constraints rather than physical constraints.
$or a manufacturing organization, #ith the goal being to make money no# as #ell as in
the future, T2+ defines three operational measurements that measure #hether operations
are #orking to#ard that goal. They are>
Throughput> The rate at #hich the system generates money through sales. This is
considered to be the same as +ontribution Margin (selling price ;; cost of ra#
materials). &abor costs are considered to be part of 2perating )xpense rather than
throughput.
0nventory> .ll the money the system invests in things it intends to (or could) sell.
This is the total system investment, #hich includes not only conventional
inventory, but also buildings, land, vehicles, plant, and e*uipment. 0t does not
include the value of labor added to :ork;0n;Process inventory.
2perating )xpense> .ll the money the system spends in turning 0nventory into
Throughput. This includes all of the money constantly poured into a system to
keep it operating, such as heat, light, scrap materials, depreciation, etc.
The follo#ing four measurements are used to identify results for the overall organization>
Net ro)it M Throughput ; 2perating )xpense
8eturn on In*estment 18OI4M (Throughput ; 2perating )xpense) 4 0nventory
roducti*it& M Throughput 4 2perating )xpense
Turno*er M Throughput 4 0nventory
'iven the measurements as described, employees can make local decisions by examining
the effect of those decisions on the organizationZs overall Throughput, 0nventory, and
2perating )xpense. . decision that results in increasing overall Throughput, decreasing
the overall 0nventory, or decreasing the overall 2perating )xpense for the firm #ill
generally be a good decision for the business.
The Theory of +onstraints does a#ay #ith much of cost accounting. 0t is clear that
application of cost accounting principles (primarily the allocation of costs in order to
F%
make decisions at the local level) leads to poor management decisions at the department
as #ell as in upper levels of the organization. 0n fact, T2+ virtually eliminates the use of
)conomic 2rder Buantities ()2B), production lot sizes, deriving product costs, setting
prices, determining productivity measures, and the use of performance incentives.
Most individuals #ill readily see the use for the Theory of +onstraints in the
improvement of production scheduling or in improving manufacturing. This is simply
incorrect. .lthough it is true that the Theory of +onstraints provides us #ith simple
examples in the manufacturing environment, T2+ is truly applicable to any process in
any organization. This includes universities, hospitals, service providers of all varieties,
government and, of course, manufacturing.
F<