Resource Optimization:

An introduction for EGMP 57

Day 2

Prof. Shubhabrata Das

IIM Bangalore

Action Plan

 Bluegrass Farms

 Manvi Motors

 Media Selection

 Resource requirement
at Station ticket
counter

Lecture notes by Prof. S. Das, IIM Bangalore 1

 Bluegrass Farms
 Bluegrass Farms has been experimenting with a special diet 
for its racehorses. The feed components available for the diet 
are a standard horse feed product, a vitamin‐enriched oat 
product and a new vitamin and mineral feed additive. The 
nutritional values in units per pound and the costs for the 
three‐feed component is given in table below:
Feed Standard feed Enriched Oat Additive
component
Ingredient A 0.8 0.2 0
Ingredient B 1 1.5 3.0
Ingredient C 0.1 0.6 2.0
Cost per pound $0.25 $0.6 $3.0

The minimum daily diet requirements for each horse are three units of
ingredients A, six units of ingredient of B and four units of ingredient C. In
addition, to control the weight of the horses, the total daily feed for the horse
should not exceed 6 pounds.

Bluegrass Farms would like to determine the optimal diet.

Ex: A minimization problem

BLUEGRASS FARMS
Minimize 0.25S  0.6 E  3 A Diet for race-horse
Feed-components:
subject to S, E and
A(vitamin, mineral additive)
0.8S  0.2 E 3 Constraints: daily
S  1.5E  3 A  6 requirements of 3
ingredients, A, B and C
0.1S  0.6 E  2 A  4 Also max weight restriction
S E A6
S  0; E  0; A0

Lecture notes by Prof. S. Das, IIM Bangalore 2

 Sensitivity analysis:
BLUEGRASS FARMS

Sensitivity analysis:
BLUEGRASS FARMS
 Slack / Surplus: the optimal diet exceeds minimum
requirement for ingredient B by 3.554 units, just meets
the requirements for ingredients A and C
 The optimal diet attains maximum permissible weight of
6 pounds
 A unit increase (decrease) in ingredients A or C
requirement would increase (decrease) in optimal cost
by 1.22 unit or 1.96 unit respectively. No impact of
small change in ingredient B requirement
 A unit increase in total feed constraint will lead to cost
reduction of 0.92 unit.

Lecture notes by Prof. S. Das, IIM Bangalore 3

 SIMULTANEOUS change: 100% rule

 100% rule for objective function coefficients: If the sum of

the percentage changes (% of allowable increase/decrease)
does not exceed 100%, the optimal solution will not change
(objective function would)

 100% rule for constraint right-hand-side: If the sum of the

percentage changes (% of allowable increase/decrease)
does not exceed 100%, then the dual prices will not change.

Manvi Motors

Lecture notes by Prof. S. Das, IIM Bangalore 4

 Formulating Manvi Motors
optimization
Decision variables: C= no of cars to produce, T= no of trucks to produce

Objective: Maximize Profit

P=(4300-800)*C + [6000 – 800(?1000)]*T - 22000

Constraints:
Fabrication hour (F) capacity (max) 12000
Assembly hour (A) capacity (max) 10000
F= 20C + 40T
A= 25C + 10 T

F+A at least 14000

C at most 500 at least 150
T at most 200 at least 100

Manvi Motors
Decision variables: C= #cars to produce T= # trucks to produce
4300 - 800

Objective function: Maximize 3500 C+5200 T -10000 - 12000

6000 - 800

Constraints:
F  20C  40T  12000
A  25C  10T  10000
F  A  14000
150  C  500
100  T  200

10

Lecture notes by Prof. S. Das, IIM Bangalore 5

 Optimal solution

Final Reduced Objective Allowable Allowable

Cell Name Value Cost Coefficient Increase Decrease

$B$3 Car DV 350 0 3500 9500 900
$B$4 Truck DV 125 0 5200 1800 3800

Final Shadow Constraint Allowable Allowable

Cell Name Value Price R.H. Side Increase Decrease

$K$6 used fabrication 12000 118.75 12000 2400 800

$L$6 used assembly 10000 45 10000 1000 3000

$N$6 used total labour 22000 0 14000 8000 1E+30

11

Discuss each what if…

1. NO. It was not optimal even under old cost structure!
2. 350 cars, 125 trucks
3. Fabrication dept (shadow price higher). But of course no. of additional hours need to be
substantially high to make any positive impact.
4. Yes. Profit would go up by 200*112.5 – 3000. If available fabrication hours increase to
13000, profit would increase by 1000*112.5 (less additional cost, if any)
5. a) 1000* 50. b) Cannot tell from the sensitivity report as 1100 increase is beyond
admissible (without changing nature of solution). By re-running, we find the optimal
solution in that case, would be to produce 400 cars and 100 trucks, yielding profit of
1878000
6. The optimal solution will not change. Profit will decrease by 500*125 $
7. No change. (intuitively, and confirmed by sensitivity analysis report)
8. No change. Optimal solution uses 2200 hours of labour.
9. a) No vans should be produced.
b) 375$ c) More than 4375 $

12

Lecture notes by Prof. S. Das, IIM Bangalore 6

 Standard LP vs IP
 In some application framework, the decision variables should be integers

 Incorporating this additional constraint, the problem is called an IP (integer

programming)

 LP –relaxation of an IP
 The original problem is an IP, but the integral-value-constraints are ignored
 If the optimal solution happens to be integral, then this solution is also the optimal
solution of the IP
 However other-wise, the “nearest” integral solution is not necessarily the optimal
solution of the IP
 GRG nonlinear method implemented by EXCEL solver is one algorithm/method to
solve IP, but the provided solution is not guaranteed to be optimal as we see in MANVI
MOTORS with constraints
 V >= 1
 V >=2

13

MEDIA Selection:
A Marketing application of LP
 Marketing manager has to allocate a fixed budget among the
various advertising media.
 The objective is to maximize reach, frequency, and quality of
exposure.
 Restrictions on the allowable allocation are driven by:
 company policy,
 contract requirements,
 budget
 media availability.

14

Lecture notes by Prof. S. Das, IIM Bangalore 7

 MEDIA Selection:

 FIVE choices of advertising media: DayTV 1min slots (DTV), Marketing

Evening TV 30sec slots (ETV), Daily Newspaper full page ad( DN),
Choice
Sunday newspaper ½ page colour (SN) & Radio
Of a
 Decision variables: Number of ads of each type Real-estate
 The objective is to maximize aggregate exposure quality project
 Restrictions:
 At least 10 TV commercials
 Reach at least 50,000 potential customers
 Budget $ 30 K (month)
 Budget for television commercials (max) $18 K
 Max ads of the 5 types: 15, 10, 25 , 4, 30

15

16

Lecture notes by Prof. S. Das, IIM Bangalore 8

 Media Selection:
Optimal solution via SOLVER

17

Media Selection: Sensitivity

analysis

18

Lecture notes by Prof. S. Das, IIM Bangalore 9

 Summary of conclusions from
sensitivity analysis: Media selection
 The exposure quality of evening TV would have to
increase by at least 65 for this media alternative to be
part of optimal selection combination
 An unit increase in budget (cost) leads to 60 units of
increase in exposure quality
 Reduction of 1 television commercial would increase
exposure quality by 25 units
 Similarly an unit increase in availability of daily
newspaper slot or radio slot would increase the
exposure quality by 16 and 14 units respectively

19

Resource requirement at
railway counter
 Decision variables
 X1= # of managers to start at 8am
 X2= # of managers to start at noon
 X3= # of managers to start at 4pm
 X4 = # of managers to start at 8pm
 X5= # of managers to start at midnight
 X6= = # of managers to start at 4 am
 Objective is to minimize = X1+X2+X3+X4+X5+X6
 Constraints X 1  X 6  12
X 1  X 2  10
X 2  X 3  12
X3 X 4  9
X4 X5 5
X5 X6 8
ALL X'S are non-neg integers
20

Lecture notes by Prof. S. Das, IIM Bangalore 10

 Resource person requirement
at Railway counters

Starting duty at
8:00 12:00 16:00 20:00 0:00 4:00

# of managers
objective 0
Total 1 1 1 1 1 1 function
min
8:00 AM – Noon 1 1 0 12
Noon – 4:00 PM 1 1 0 10
4:00 PM – 8:00 PM 1 1 0 12
8:00 PM – Midnight 1 1 0 9
Midnight – 4:00 AM 1 1 0 5
4:00 AM - 8:00 AM 1 1 0 8

21

MANY OPTIMAL SOLUTIONS!

++++++

Find “best” based on additional reasonable criteria

22

Lecture notes by Prof. S. Das, IIM Bangalore 11

 Home work
 Come prepared with reading the cases (try to frame)
 Chip Hoose wheels
 Transhipment problem for Nimo electronics
 Try to solve Navy shipping problem (given next – will be discussed
in the next class)
 Solve exercises from the textbook, focusing on
 LP / IP formulation
 Framing in excel and finding optimal solution via SOLVER
 Understand & Interpret sensitivity analysis

23

Transportation Problem: Example #2

Navy Shipping

The Navy has 9,000 tons of material in Vizag,

that it wishes to ship to three installations:
Mumbai, Chennai, and Kolkata. They require 4,000,
2,500, and 2,500 tons, respectively. Government
regulations require equal distribution of shipping
among the three carriers.

24

Lecture notes by Prof. S. Das, IIM Bangalore 12

 Transportation Problem: Example #2
The shipping costs per pound for truck, railroad,
and airplane transit are shown below.
Formulate and solve a linear program to determine the
shipping arrangements (mode, destination, and
quantity) that will minimize the total shipping cost.

Destination
Mode Mumbai Chennai Kolkata
Truck 12 6 5
Railroad 20 11 9
Airplane 30 26 28

25

Lecture notes by Prof. S. Das, IIM Bangalore 13