536 Int. J. Logistics Systems and Management, Vol. 20, No.

4, 2015

Lean practices in machinery manufacturing

industries – a case study

P.G. Saleeshya*
Department of Mechanical Engineering,
Amrita University,
Coimbatore, 641007, India
Email: pg_saleeshya@cb.amrita.edu
*Corresponding author

A. Sneha
Caterpillar India Pvt. Ltd – Engineering Design Center,
Ascendas IT Park 7th Floor,
Tharamani Road, Chennai – 600113, India
Email: sneha_1992@hotmail.co.uk

C. Karthikeyan
Tata Technologies Limited,
25 Rajiv Gandhi Infotech Park,
Hinjawadi, Maharashtra Pune 411057, India
Email: karthikeyan29692@gmail.com

C. Sreenu
Ashok Leyland Ltd. MDC,
312, Sipcot Industrial Complex,
Hosur – 635 126, Tamil Nadu, India
Email: sreenu.s5@gmail.com

A.K. Rohith
Sundram Fasteners Limited,
Harita, Hosur 635109, India
Email: rohithkoroth@gmail.com

Abstract: Lean manufacturing is an emerging concept in the Indian

manufacturing scenario. Many small and medium scale enterprises in the
country are aiming at becoming lean and thereby improving their productivity.
This paper is an outcome of a case study in one such enterprise that deals with
the manufacturing and assembly of offset printing machinery in India. Being a
vendor-based organisation, part shortage due to delays in the process was a
major problem faced. A thorough analysis of the various vendors and processes
that the product goes through was done and the process flow was mapped to
obtain a clear picture of the entire system, and software modelling helped arrive

Copyright © 2015 Inderscience Enterprises Ltd.

 Lean practices in machinery manufacturing industries – a case study 537

at quantitative results. Lean tools like Pareto analysis, five-why analysis, source
inspection, etc., were used to identify and eliminate the various factors that led
to delay. Proper scheduling of jobs is also essential to reduce waiting time, and
therefore an algorithm has been developed to help arrive at an optimum
scheduling sequence. The results from this case study have proved successful in
reducing the lead time of the component under study.

Keywords: lean manufacturing; LM; value stream mapping; VSM; lead time;
genetic algorithm; supply chain agility; logistics.

Reference to this paper should be made as follows: Saleeshya, P.G., Sneha, A.,
Karthikeyan, C., Sreenu, C. and Rohith, A.K. (2015) ‘Lean practices in
machinery manufacturing industries – a case study’, Int. J. Logistics Systems
and Management, Vol. 20, No. 4, pp.536–554.

Biographical notes: P.G. Saleeshya is a Professor of Mechanical Engineering

at the Amrita University, Coimbatore, India. She received her PhD in Industrial
Engineering and Operations Research (Agile Manufacturing) from Indian
Institute of Technology, Bombay, India. Her areas of interest in teaching and
research is mainly focused on operations strategy: agile manufacturing, lean
manufacturing, Lean Six Sigma, green design and manufacturing, and
responsive supply chain, agility in textile industries. She has several
publications in various international journals.

A. Sneha graduated from Amrita School of Engineering with Mechanical

Engineering as the main stream. She is currently working as an Associate
Engineer with Caterpillar India Pvt. Ltd. in Chennai. Her fields of interest are
lean manufacturing and supply chain management and this is the reason she
chose to work in a core manufacturing sector. She aspires to pursue her masters
in the same domain.

C. Karthikeyan is a graduate student in the Department of Mechanical

Engineering at the Amrita School of Engineering Coimbatore, Tamil Nadu.
Currently, he is a graduate engineering trainee in engineering design at the Tata
Technologies, Pune. His research interests are in lean manufacturing,
automobile design and manufacturing, advance manufacturing technologies,
supply chain management and logistics management.

C. Sreenu did his graduation in Mechanical Engineering from Amrita School of

Engineering, Coimbatore. He is currently a graduate engineering trainee with
Ashok Leyland Ltd. His fields of interest are lean manufacturing and
mathematical modelling and analysis.

A.K. Rohith is a graduate student in the Department of Mechanical Engineering

from Amrita Vishwa Vidyapeetham University, Coimbatore. He is currently
working as a graduate engineer trainee at the TVS Sundram Fasteners Limited,
Hosur, Tamilnadu. His areas of interest include lean manufacturing process
optimisation and productivity improvement.

1 Introduction

Lean is a relatively modern management concept which aims at eliminating wastes and
increasing efficiency. The term ‘lean’ denotes a system that utilises less, in terms of all
538 P.G. Saleeshya et al.

inputs, to create the same outputs as those created by a traditional mass production
system, while contributing increased varieties for the end customer (Womack and Jones,
2003). The resounding principle of lean manufacturing (LM) is to reduce cost through
continuous improvement that will eventually reduce the cost of services and products,
thus growing more profits (Liker, 2004). Lean principles mainly aim at eliminating waste
in every process activity to reduce process cycle, improve quality and increase efficiency.
In lean context, waste (Muda in Japanese) includes all forms of overproduction,
over-processing, unnecessary transport, excess motion, defects, waiting and inventory
(Ohno, 1988; Shingo, 1989).
To achieve leanness, a set of lean tools and techniques are available such as Kanban,
total preventive maintenance (TPM), quality at the source, Pareto analysis, five-why
analysis and Kaizen. The objective of this paper is to use such lean tools and techniques
to improve the process flow of a critical component of printing machinery, thereby
reducing its lead time. Also, since scheduling can influence the time factor to a great
extent, genetic algorithm (GA) (an evolutionary algorithm) has been used to help arrive at
an optimum scheduling sequence and it helps reduce the time spent by a job waiting to be
processed.
All the above contribute to achieving a leaner future state when compared to the
current state of the system. Zapp et al. (2013) used lean knowledge management systems
on the shop floor to identify redundant work and poor product quality, which is prevented
by the effective use of workers knowledge. Rathje et al. (2009) considers lean success to
occur if a company achieves the major strategic components of lean such as management
commitment, employ autonomy, information transparency and cultural fit, and
successfully implements a number of practices to support the operational and tactical
aspects of lean. Many companies are implementing lean with the objective of achieving a
superior competitive advantage over other companies. Few of them have attained their
objectives, while many of them have not. One of the reasons for failure is that the
managers of these organisations have not understood clearly how the performance
measures of each organisation are affected when it gets transformed through LM. To fill
this gap, Anand and Kodali (2009) developed a simulation model-QUEST Siml, which
will help the managers analyse the impact of various lean elements on the performance
measures of the organisation.
Mass, lean and agile production philosophies, although widely discussed, still cause
considerable confusion among both in academics and in practice. Kisperska-Moron and
de Haan (2010) over booms characterisations of lean (what, when needed but perfect) and
agile (first, fast and best) show the paradigmatic differences between the two.
Understanding of agile framework as well as lean methodologies and their
implementation methods are widely different from organisation to organisation
(Saleeshya et al., 2012b; 2013). Welo et al. (2013) used lean system engineering (LSE) to
improve systems through engineering design improvement.
The remainder of the paper is structured as follows: Section 2 reviews the related
literature, Section 3 presents the preliminary study of the system with modelling and
simulation of the current state, Section 4 presents the problems identified and the
methods to tackle them using various lean tools, Section 5 involves the application of GA
to solve the scheduling problem, Section 6 presents the future state of the system after
application of lean tools and scheduling, Section 7 presents conclusions and insights, and
Section 8 presents the future research scope.
 Lean practices in machinery manufacturing industries – a case study 539

2 Literature review

Original lean thinking was closely associated with the Toyota Production System and the
work of Ohno (1988), focusing on eliminating excess, waste and unevenness. The
superior performance of LM systems has encouraged the idea of transferring LM to non-
Japanese and non-automotive situations, with results proving the exemplary nature of this
concept.
Lamming (1996), generalised the concept of lean to encompass other industries and
extended its application from manufacturing firms to supply chain management. A lean
supply chain as defined by Lamming was “... an arrangement [which] should provide a
flow of goods, services and technology from supplier to customer (with associated flows
of information and other communications in both directions) without waste”. Now, lean
thinking not only has been applied in manufacturing industries, but also is being adopted
by service industries (Pool et al., 2010). Gorane and Kant (2014) identified supply chain
practices (SCPs) as the most important strategies for medium and large organisations and
has become a prerequisite for success in the global market (Wisner and Stanley, 2008).
SCPs act as important tools to achieve competitive advantage together with strengthening
organisational competitiveness.
Pool et al. (2011) in their study consider how the principle of ‘flow’ and ‘pull’
production-suggesting a regular, demand-driven product flow may be implemented for
the semi-process industry by introducing cycle schedules. The lean concept embraces a
wide range of activities and tools within and beyond an organisation’s boundary. Loader
(2010) makes a contribution to the research debate examining relationships between
public and private sector procurement practices in UK local authorities (Hines et al.,
2008). Salleh et al. (2012) developed a simulation model using Delmia Quest Software
and used lean kaizen to improve flow by increasing productivity and reducing working in
progress. Steinberg and De Tomi (2010) modelled a supply chain to identify and
understand the points of value creation and places for possible improvement
interventions.
Mothersell et al. (2008) argued that the change to lean also identifies and discusses
culture and people systems in harmony with lean and parallel organisational structures.
According to Michael Gnanaraj et al. (2009), Lean Six Sigma is tailor-made to suit large
size organisations. There has been little effort on implementing this paradigm in small
and medium enterprises and this will help large organisations to achieve global
competitiveness. The research by Dombrowski et al. (2012) showed that a multitude of
different knowledge flows can occur during the implementation of lean production
systems and that a decentralised, role-specific approach can help to identify adequate
methods of knowledge management.
Gnoni et al. (2013) proposes innovative design of a near-miss management system
(NMS) based on the integration of principles of lean management in occupational safety
for a worldwide automotive supplier firm. Salleh et al. (2012) proposed a green lean total
quality (GLTQ) information management system that comprises of information
management (IM) in environmental management system (EMS) practices which is
integrated to TQM with LM principles. This proposed green LTQM IM model can be
conducted to access the status of IM in any company. Nivrithi et al. (2013) have
improved process capability to increase productivity by reducing cycle time.
540 P.G. Saleeshya et al.

All these studies reported in the literature motivated the authors to conduct the
present study in a machinery manufacturing company in which the authors have tried to
implement various lean tools, the details of which are presented in the following section.

2.1 Overview of lean tools and concepts

A very brief description of the most common lean tools and concepts are given below.
The interested reader can refer to one of the many books on LM for more details
(Monden, 1998; Feld, 2000).
• Pull production: Regulates the flow of resources in a manufacturing process by
replacing only what has been consumed and only what is immediately deliverable.
Customer order drives the production schedules based on actual demand and
consumption rather than forecasting.

• Kanban: A signalling system for implementing pull production. It is essentially a

work order that also moves with the material. Each card or Kanban identifies the part
or subassembly unit and indicates where each came from and where each is going.

• Poka-yoke: It helps people and processes work right the first time. Poka-yoke refers
to the technique that makes it impossible to make mistakes thereby helping to drive
defects out of products and processes.

• TPM: Workers carry out regular equipment maintenance to detect any anomalies.
The focus is changed from fixing breakdowns to preventing them. Since operators
are the closest to the machines, they are included in maintenance and monitoring
activities in order to prevent and provide warning of malfunctions.

• 5S: Focuses on effective work place organisation and standardised work procedures.

• Cause-and-effect diagram: Also, known as fish-bone diagram or Ishikawa diagram, it

is a tool that helps identify, sort, and display possible causes of a specific problem or
quality characteristic. It graphically illustrates the relationship between a given
outcome and all the factors that influence the outcome.

• Pareto analysis: It is a statistical technique in decision making that is used for

selection of a limited number of tasks that produce significant overall effect.

• 5-why analysis: It is an iterative question-asking technique used to explore the cause-

and-effect relationships underlying a particular problem. The primary goal of this
technique is to determine the root cause of a defect or problem.

2.2 Job scheduling and overview of GAs

Scheduling is a decision making function to allocate resources over time to perform a
collection of tasks. Scheduling problems are classified into different categories based on
the shop configuration, solution techniques, certainty of data, and arrival of jobs to the
system. They can be classified as single machine, parallel machine, flow shop, job shop,
flexible manufacturing, cellular manufacturing, and assembly line balancing in terms of
shop configuration.
 Lean practices in machinery manufacturing industries – a case study 541

If the information about the jobs is available with certainty, those problems are called
deterministic. Even if only one parameter is not known with certainty, the problem is
referred to as stochastic. In some cases, all the jobs may be available at time zero (static)
and in others they may be arriving individually at different times (dynamic). Solution
procedures are classified as optimising or heuristic procedures. Most of the optimising
procedures employ branch-and-bound and mathematical modelling approaches which are
still computationally demanding (Deogun, 1983). As a result, a lot of research is going on
in developing fast and efficient heuristic procedures.
One such heuristic procedure is GA, which is a search method that mimics the natural
biological evolution (Gen and Cheng, 1996). GA operates on a population of potential
solutions applying the principle of survival of the fittest to produce better approximations
to a solution. At each generation, a new set of approximations is created by the process of
selecting individuals according to their level of fitness in the problem domain and
breeding them together using operators such as crossover reproduction and mutation
borrowed from natural genetics. This process leads to the evolution of populations of
individuals that are better suited to their environment than the individuals that they were
created from, just as in natural adaptation.

3 Preliminary study of the system

The company under study is largely vendor-based and even though this offers huge
advantages, their main challenges also arise due to the same. Delay in supply of parts,
caused mainly due to prioritisation and scheduling issues at the vendor’s end, has a huge
impact on the time to assemble the printing machinery and deliver it to its customers.
Also, since there are some parts that are outsourced multiple times to different vendors
before it becomes the final product that can be used for assembly, inspections at each
stage consume quite a large amount of time. Quality rejection of parts is another problem
in hand. The more the number of parts rejected, the more delay caused and the more time
the assembly plant is waiting for parts to arrive.
Our paper focuses on improving the system of production of parts for assembling
through lean for lead time reduction of the plate cylinder of a printing machine which has
the highest customer demand. We expect to achieve this objective through analysis of the
current state condition and suggesting lean and kaizen solutions to the problems
identified. This plate cylinder is critical because it goes through multiple vendors before
reaching the final stage. Also, dimensional tolerance and precision become major
necessities as printing cylinders are the major components that determine the quality of
print. So, improving and standardising the process of plate cylinders will help the
company to be more productive and competitive. The first step would be to identify the
problems faced by the system through value stream mapping (VSM).

3.1 Analysis of current condition of the system

Being largely a vendor-based company, its overall supply chain is widely spread out
(Abdulmaleka and Rajgopal, 2006). Therefore, the lean tool VSM is used to help look at
the entire process flow as one ‘system’. A value stream is a collection of all actions
(value-added as well as non-value-added) that are required to bring a product (or a group
542 P.G. Saleeshya et al.

of products that use the same resource) through the main flows, starting with raw material
and ending with the customer (Rother and Shook, 1999). The current condition value
stream map is provided below as shown in Figure 1.

Figure 1 Current condition value stream map (see online version for colours)

It is seen that the lead time taken by a cylinder to go through all the processes is 56 days.
While actually, the value added time, that is the time an operation actually takes place on
the product, is a mere 1.28 days. This reveals the huge scope of improvement that’s
possible in the system. Around 54 days is the non-value added time that has to be reduced
as much as possible. Now, to understand the situation more clearly, the next step is going
to be a simulation of the current scenario which will give us more quantitative
information that can be worked on.
Table 1 Current state simulation results

VA time (avg) NVA TIME (avg) Number out

Casting 13.5 hrs 12.5 days 26
Pre-machining 1.5 hrs 15.8 days 25
CNC 3.5 hrs 6.7 days 24
Pre-grinding 1.7 hrs 7.1 days 21
Coating 8 hrs 7.8 days 17
Post-grinding 2.67 hrs 4.2 days 16
Total 1.286 days 54.1 days
 Lean practices in machinery manufacturing industries – a case study 543

Table 1 shows the results obtained from running the simulation. The simulation results
highlighted that the number of cylinders that can be completed during a period of
56 days, is 16. Whereas, based on forecast, the required number of cylinders per month
would be 16. Which means the current situation is able to satisfy only half of the real
demand.

3.2 Problems identified and solutions proposed

From the results obtained in the previous section, we were able to identify the major
problems occurring within the system. They are as listed below:
• Inconsistent order placement times and batch sizes resulting in delays in the process
due to part shortage. The batch sizes vary from a minimum of 1 to a maximum of
36 cylinders. The time between order placements are also found to vary from a
minimum of 1 to a maximum of 54 days.
• Most of the lead time extension is caused due to delays in the process flow. This has
to be majorly accounted to the priority rules that the vendors use. The product is
found waiting for more than 80% of the lead time.
• On analysis of rejected materials, it is found that most of the rejections occur due to
casting defects. Unfortunately, these defects can be identified only at the later stages
when a lot of value has already been added to the product. When the product itself is
so critical and takes so much of time for completion, rejection at the last stages is a
major consumer of time, cost and effort.
• A major problem in India these days has been the power cuts and that has an effect
on most industries. A lot of the company’s and its vendor’s capacity is restricted due
to power shortage. A solution to this issue is out of our scope.

3.3 Stabilising the current condition

According to 5s, we have to sort down the information, streamline the current flow and
make it stable before improving and standardising it. Now, that the problems have been
identified and quantified in the process flow, the next step would be to look at how
improvements can be brought about to the process, which will reduce the time and in turn
the cost of manufacturing the cylinder. But before we look at improvements, we would
like to propose a system that would make sure that in the current scenario, the company
does not face shortage of this critical component. This is by helping the company
understand when to place its orders and how much quantity to order using a
determination of order placement time model.
It should be noted that each vendor is considered as a separate system and the model
is applied to them individually. This is because each individual vendor has its own
inventory levels and overall costs, which would make it more economical for them than
when treated as a whole. At the outset, any individual cost reductions will ultimately
result in the reduction in total cost of the entire system too. The results obtained are
tabulated in Table 2.
544 P.G. Saleeshya et al.

Table 2 Economic order quantity, reorder point and reorder interval for process

Problem Root cause Solution

Blow holes Low permeability of sand due Reduce moisture content in sand.
to high moisture content Increase number of vent holes.
Reduce ramming.
Porosity Insufficient pre-heating of mould Pre-heat the mould to a sufficient
extent to allow gases to escape.
Peel off Inadequate cleaning before plating Cleaning with organic solvent, acid
pickling and pre-heating.
White dots Interruption of electric Use a modern three phase rectifier to
current while plating convert AC to DC smoothly.

From the results of the above model, it is evident that there will be some inventory stock
up at the end of each process stage since the order quantities and order placement times
vary from process to process. Though this might contribute to a slight hike in inventory
carrying cost, it is going to ensure the availability of the cylinder when in need at the
assembly floor. Since the lead times for each process include the delays and other non-
value added times, this model’s input data is real time and hence can be used effectively
to understand when to place and what quantity of each order to place to avoid part
shortage.
Now, that a stable system has been established, in the next section we look at how
certain lean methodologies and tools can be employed to identify and eliminate waste in
the process.

4 Application of lean tools and techniques

4.1 Lean methodology

The product under focus and major problems faced were identified in the previous
section. Now, various lean tools that could be applied to identify and eliminated waste
from the system are discussed.

4.1.1 Cause and effect diagram

The major problem in the process flow of the plate cylinder was found to be delay. So, an
exhaustive list of the various factors causing delay is presented in Figure 2.
Now, that the major factors contributing to delay have been identified, the highlighted
problems are aimed at being tackled. The factors causing delay that could be eliminated
were chosen based on the feasibility and practicality of an improved solution.
 Lean practices in machinery manufacturing industries – a case study 545

Figure 2 Cause and effect diagram for delay (see online version for colours)

4.1.2 Vendor relationship

Extended organisation is an important concept in any lean implementation, more so for a
company that handles so many vendors. Transparency must be ensured in the system and
mutual trust and commitment must be developed. For any lean implementation to work
the whole system must commit to becoming lean and this is impossible without the
cooperation of the vendors. A vendor development team is established in the company to
rate the vendors based on quality, the time of delivery and the quantity of product
delivered. So, the vendors who rank high in delivering the right amount with proper
quality at the right time were considered and monitored periodically to have an extended
organisation with transparency.

4.1.3 Quality rejections and rework

Pareto analysis has been used to analyse the major factors contributing to 80% of quality
rejections and rework. Quality rejections occur majorly due to casting defects and quality
rework occurs mostly due to plating defects. On collecting relevant data, Pareto charts
have been plotted as represented in Figures 3 and 4.
546 P.G. Saleeshya et al.

Figure 3 Pareto chart for quality rejections

Figure 4 Pareto chart for quality rework

On identifying the major factors, kaizen problem solving sheets were used to identify the
root cause of each problem and its corresponding counter-measure. The findings have
been tabulated in Table 3.
Table 3 Counter-measures for major quality rejections and rework

Reorder interval (T*)

Process Quantity (Q*) Reorder point (r*)
[days]
Casting 18 7 33.4
Pre-machining 16 8 30.4
CNC milling 14 4 26.8
Pre-grinding 14 4 26.5
Hard chrome plating 12 4 23.5
Post-grinding 12 2 23.3
 Lean practices in machinery manufacturing industries – a case study 547

4.1.4 CNC downtimes

CNC downtimes were also found to be a major cause for delay according to the fish bone
diagram charted. Pareto analysis was used and the major factors contributing to
downtime were found to be operator not available, no load, power cut and machine
cleaning. Operator not available occurs when the operators are breaking for lunch time or
meetings. Modification of shift timings such that they overlap during the time consumed
by lunch can be done so that the machine is utilised effectively.
To prevent no load condition, proper load scheduling must be done so that the
machine is neither completely occupied by queued parts nor is it starving for components.
A solution to power cuts is outside the scope of this paper. Machine cleaning can be
avoided if TPM is meticulously followed. Though the company has a TPM checklist, the
practice is not being followed on a regular basis. A 5S survey was conducted and the
company was found to implement only the first 3Ss. Suggestions to improve the first 3Ss
and also incorporate the next 2Ss were provided.

4.2 Kaizen
4.2.1 Source inspection
During the system study, it was identified that inspections happened multiple times. It
was done once by the quality control department and once by the machine operator
before and after machining. To avoid repetitive measurement of the same dimension, the
use of a dimension tag was suggested. It is a matrix that contains the standard dimensions
required at the end of each process and a space to note down the measured dimensions at
each stage. This tag travels along with the cylinder. The dimension tag not only
eliminates the need for multiple checks but also visually portrays any deviation between
the standard and observed dimension. On implementing source inspection, the calculated
reduction in delay is 31.2% by making inspection an integrated part of the process.

4.2.2 Setup time reduction

CNC setup time reduction was carried out using an ECRS sheet as shown in Figure 7.
ECRS stands for eliminate, combine, restructure and simplify. A detailed study of the
various activities that the CNC operator carries out during machine setup and changeover
were noted. Activities which could be carried out at an earlier time were identified and
eliminated from the setup procedure. Some activities were also simplified to reduce the
time consumed in that process. The setup time was reduced from 100 minutes to
71 minutes and the changeover time was reduced from 5 minutes to 1 minute.
The next major problem which is the one that has not been tackled using lean tools is
scheduling. This will be dealt with in the next section.

5 GA to solve scheduling problem

This section focuses on single machine scheduling for n jobs with non-zero ready time
with an objective to sequence jobs such that the average flow time is minimised. These n
jobs are non-preemptible, independent and have equal priority of execution.
548 P.G. Saleeshya et al.

When preemption is not allowed, the scheduling task gets more complex. If the next
job in the sequence has ready time lower than the current system time, it is assigned to
the machine right away. On the other hand, if the ready time of the next job in the
sequence is greater than the current system time, then there are two possible strategies:

• keep the machine idle until the job becomes available (delay strategy)

• keep the machine busy by assigning the next available job in the sequence
(non-delay strategy).

The current average flow time as observed in the company is 3.7719 days.

5.1 GA reproduction parameters

The different reproduction parameters or genetic operators used in solving the GA are
given below:
• Mutation: Mutation is a genetic operator used to maintain genetic diversity from one
generation of a population of GA chromosomes to the next. It is analogous to
biological mutation. A mutation probability of 0.2 is used. Based on the mutation
probability, the level of mutation that has to be done in a particular chromosome is
found out. The genes which can undergo mutation are then swapped among
themselves at random.
• Crossover: Crossover is a genetic operator used to vary the programming of a
chromosome or chromosomes from one generation to the next. It is analogous to
reproduction and biological crossover, upon which GAs are based. Cross over is a
process of taking more than one parent solutions and producing a child solution from
them. A crossover probability of 0.8 is used. A new method is developed to carry out
crossover so as to avoid the problem of infeasibility and which also effectively
retains the properties of both parents is explained in the next sub-section.

5.1.2 Proposed crossover method

Let each parent be represented by a permutation of ‘n’ unique numbers that represent the
n jobs in the sequence. For example, two different sequences of five jobs each is
represented as X1 = [1 2 3 4 5] and X2 = [2 1 5 3 4]. After a two point crossover, with
start point as position 2 and end point as position 3, one of the two offspring formed
would have the sequence represented as [2 2 3 3 4]. As it is evident that job 2 and job 3
are repeating in the offspring, this becomes an infeasible sequence.
To avoid this problem of job repetition after crossover, many methods are available
from literature, the most popular of them being partially matched crossover (Goldberg
and Lingle, 1985), edge recombination (Whitley et al., 1989), ARGOT strategy (Shaefer
and Smith, 1990) and random keys (Bean, 1993). A continuing drawback of these
methods has been the need for specialised representation for each problem variation, an
exception being random keys. But, on analysing this method it was found out that the
properties of parents were being ignored by offspring to a large extent which contradicts
the basic idea of crossover. Therefore, a new crossover method is proposed as follows:
 Lean practices in machinery manufacturing industries – a case study 549

1 Two parent chromosomes X1 and X2 are chosen from current population. E.g.,
X1 = [1 2 3 4 5] and X2 = [2 1 5 3 4].
2 Each gene in X2 is multiplied by a common factor p (for e.g., 100) before crossover.
Now, X2 = [200 100 500 300 400].
3 Let us assume two point crossover is being performed between two random points
(for e.g., with start point as position 2 and end point as position 3). One of the
resultant offspring is [200 2 3 300 400].
4 Now, each gene from this resultant offspring, excluding the genes from the crossover
start point and end point [2 3], that is [200 300 400] is divided by the common factor
p and compared for equality relationship with genes in the crossover region.
If equality relationship is satisfied, the gene remains the same in the offspring and the
position of these genes in offspring is noted down in an array P1. If the equality
relationship is not satisfied then the corresponding gene is divided by the common factor
p producing the modified offspring as [200 2 3 300 4]. P1 = [1 4].
1 Then the missing jobs in the modified offspring are found. In our case they are 1 and
5.
2 Now, these missing jobs, randomly replace the genes whose positions are stored in
P1 to produce the new offspring. For example the new offspring could be [1 2 3 5 4]
or [5 2 3 1 4].
From the above proposed method, both infeasibility of solution is avoided and properties
of parents are retained resulting in faster convergence to optimum solution as shown in
Figure 5.

Figure 5 Best fitness value graph (see online version for colours)

5.2 Results obtained from GA application

The average flow time obtained from best sequence was found to be 0.6638 days, when
compared to initial average flow time of 3.7719 days from computational results as
shown in Figure 5.
550 P.G. Saleeshya et al.

The application of proper scheduling has resulted in eliminating time wastage to a

large extent, as much as 82% reduction in average flow time of a part, and so this
becomes a useful tool that can be used by the company to schedule their jobs.

6 Future state VSM and simulation of the system

Based on the findings our study as reported in Sections 4 and 5 the following changes
have been made to the current condition map and simulated the entire system, to develop
the future state as shown in Figure 6.
• The order placement quantity and time interval were given according to the
‘determination of order placement time’ model.
• Quality rejections and rework were removed from the system. Since 80% of the
problems causing these rejections have been eliminated through Pareto analysis, the
rest of the 20% will be negligible and can be ignored.
• Inspection times were removed as a consequence of applying the system of source
inspection.
• The setup times for CNC and grinding machine were reduced as per the suggestions
proposed in the work analysis sheet.
• The waiting time before the CNC machine was reduced to 0.66 days which was the
result of applying the concept of GA to arrive at an optimum schedule.

Figure 6 Future state value stream map (see online version for colours)
 Lean practices in machinery manufacturing industries – a case study 551

Figure 7 ECRS sheet for CNC setup time reduction

The future state value stream map is presented in Figure 6 and it is to be noted that the
lead time has decreased to a considerable level. That is, from 56 to 33.8 days. After
making the same changes in the simulation model, the model was run and the output was
found to be 16 cylinders in 34 days compared to the previous model output of 16
cylinders in 56 days. The detail of which is presented in Table 4.
Table 4 Future state simulation results

VA time (avg) NVA time (avg) Number out

Casting 13.5 hrs 8 days 21
Pre-machining 1.5 hrs 8 days 20
CNC 3 hrs 4.6 days 20
Pre-grinding 1.4 hrs 5 days 20
Coating 8 hrs 4 days 19
Post-grinding 2.17 hrs 3 days 16
Total 1.23 days 32.6 days

This implies that if the changes that are proposed are implemented, the system
performance will improve to a great extent and the demand will be satisfied. Every lean
tool and every model used to analyse and eliminate waste from the system, has
quantitatively contributed to time reduction and that effect is highlighted in the results
obtained.

7 Conclusions

The lead time reduction of the critical component, that is, the plate cylinder has been
achieved through the identification and elimination of various kinds of wastes with the
help of lean tools to identify waste and defects, process improvement (kaizen) to
eliminate error and non-value addition process, and genetic algorithm to improve
scheduling of single machine. The lead time of the cylinder was found to reduce from 56
days to 33.8 days. Also, the simulation results proved that the system, that is the result of
our findings, is capable enough to produce 16 cylinders per month and hence satisfy the
demand. This paper can be used as a model for lean implementation in similar
manufacturing industries.
A limitation of the study is that the utilisation of various resources involved in the
processing of the plate cylinder could not be analysed. Also, the paper focuses only on a
single component that the company handles, and only single objective optimisation of
schedule has been carried out. The demand has been considered to be constant according
to company forecast. The dynamic behaviour of the market has not been taken into
account.
552 P.G. Saleeshya et al.

7 Future research scope

The study could be extended to the vendors in a more detailed manner, in that, their
products can be taken into account and the various resource utilisations and effectiveness
can be evaluated and measures can be taken to improve the all resources in process using
GA form multiple resources so that the lead time would further decrease due to increased
productivity. The study can also be extended to all the components that the company
handles so that the entire system improves as a whole. The scheduling model could be
extended to handle multiple resource problems with multiple objectives and the order
placement time determined, should be able to handle the variations in market demand.
Other statistical optimisation techniques can be included and compared by which the
search for a more optimal solution can be improved.

Acknowledgements

The authors acknowledge all the unknown referees, editor of the journal and all well
wishers.

References
Abdulmaleka, F.A. and Rajgopal, J. (2006) ‘Analyzing the benefits of lean manufacturing and
value stream mapping via simulation: a process sector case study’, International Journal of
Production Economics, May 2007, Vol. 107, No. 1, pp.223–236.
Anand, G. and Kodali, R. (2009) ‘Simulation model for the design of lean manufacturing systems –
a case study’, International Journal of Productivity and Quality Management, Vol. 4,
Nos. 5–6, pp.691–714.
Bean, J.C. (1993) Genetics and Random Keys for Sequencing and Optimization, Technical Report
92-43, University of Michigan, Ann Arbor, MI.
Deming, W.E. (1982) Quality, Productivity, and Competitive Position, Massachusetts Institute of
Technology, Centre for Advance Engineering Study, pp.108–109.
Deogun, J.S. (1983) ‘On scheduling with ready to minimize mean flow time’, The Computer
Journal, Vol. 26, No. 4, pp.320–328.
Dombrowski, U., Mielke, T. and Engel, C. (2012) ‘Knowledge management in lean production
systems’, 45th CIRP Conference on Manufacturing Systems, Procedia CIRP 3, pp.436–441.
Feld, W.M. (2000) Lean Manufacturing: Tools, Techniques, and How to Use Them, The St. Lucie
Press, London.
Gen, M. and Cheng, R. (1996) Genetic Algorithm and Engineering Design, Wiley Publications,
New York.
Gnoni, M.G., Andriulo, S., Maggo, G. and Nardone, P. (2013) ‘Lean occupational safety:
an application for a near-miss management system design’, Safety Science, Vol. 53, Null,
pp.96–104.
Goldberg, D.E. and Lingle, R. Jr. (1985) ‘Alleles, Loci, and the traveling salesman problem’,
Proceeding of the First International Conference on Genetic Algorithms.
Gorane, S.J. and Kant, R. (2014) ‘Supply chain practices: study and framework for future
development’, Int. J. of Logistics Systems and Management, Vol. 17, No. 1, pp.83–103.
Hines, P., Martins, A.L. and Beale, J. (2008) ‘Testing the boundaries of lean thinking: observation
from the legal public sector’, Public Money and Measurement, Vol. 28, No. 1, pp.35–40.
 Lean practices in machinery manufacturing industries – a case study 553

Kisperska-Moron, D. and de Haan, J. (2010) ‘Improving supply chain performance to satisfy final
customers: legile experience of a polish distributor’, International Journal of Production
Economics, Vol. 125, No. 2, pp.251–261.
Lamming, R.C. (1996) ‘Squaring lean supply with supply chain management’, International
Journal of Operations and Production Management, Vol. 16, No. 2, pp.183–196.
Liker, J. (2004) The Toyota Way: 14 Management Principles from the World’s Greatest
Manufacturer, 1st ed., McGraw-Hill, USA.
Loader, K. (2010) ‘Is local authority procurement lean? An exploration to determine if ‘lean’ can
provide a useful explanation of practice’, Journal of Purchasing and Supply Chain
Management, Vol. 16, No. 2, pp.41–50.
Michael Gnanaraj, S., Devadasan, S.R. and Shalij, P.R. (2009) ‘Current state maps on the
Implementation of Lean and Six Sigma paradigms and an exclusive model for deploying Lean
Six Sigma in SME’s’, International Journal of Productivity and Quality Management, Vol. 5,
No. 3, pp.286–309.
Monden, Y. (1998) Toyota Production System – An Integrated Approach to Just-In-Time, 3rd ed.,
Engineering & Management Press, Norcross, Georgia.
Mothersell, W.M., Moore, M.L. and Strolle, M. (2008) ‘A brown field lean conversion: a case
study of Opel Belbium’, International Journal of Productivity and Quality Management,
Vol. 3, No. 2, pp.161–182.
Nivrithi, B., Devadasan, S.R., Pramodh, V., Kamlesh, M. and Praveen, B. (2013) ‘Process
improvement through capability analysis: a case study in an engine valve manufacturing
company’, Int. J. of Logistics Systems and Management, Vol. 16, No. 3, pp.244–268.
Ohno, T. (1988) The Toyota Production System: Beyond-Large Scale Production, Productivity
Press, Portland, OR.
Pool, A., Wijngaard, J. and van der Zee, D-J. (2010) ‘Lean planning in the semi-process industry, a
case study’, International Journal of Production Economics, Vol. 26, No. 1, pp.1–6.
Pool, A., Wijngaard, J. and van der Zee, D-J. (2011) ‘Lean planning in the semi-process industry, a
case study’, International Journal of Production Economics, May 2011, Vol. 131, No. 1,
pp.194–203.
Rathje, M.S., Bogle, T.A. and Deflorin, P. (2009) ‘Lean take two! Deflections from the second
attempt at lean implementation’, Business Horizons, Vol. 52, No. 1, pp.79–88.
Rother, M. and Shook, J. (1999) Learning to See: Value Stream Mapping to Add Value and
Eliminate Muda, The Lean Enterprise Institute, Inc., Brookline, MA.
Saleeshya, P.G., Austin, D. and Vamsi, N. (2013) ‘A model to assess the lean capabilities of
automotive industries’, Int. J. Productivity and Quality Management, Vol. 11, No. 2,
pp.195–211.
Saleeshya, P.G., Raghuram, P. and Vamsi, N. (2012a) ‘Lean manufacturing practices in textile
industries – a case study’, Int. J. Collaborative Enterprise, Vol. 3, No. 1, pp.18–37.
Saleeshya, P.G., Thampi, K.S. and Raghuram, P. (2012b) ‘A combined AHP and ISM based model
to assess the agility of supply chain – a case study’, International Journal of Integrated Supply
Chain, Vol. 7, Nos. 1/2/3, pp.167–198.
Salleh, N.A.M., Kasolang, S. and Jaffar, A. (2012) ‘Green lean total quality information
management in Malaysian automotive company’, International Symposium on Robotics and
Intelligent Sensors (IRIS), Procedia Engineers, Vol. 41, pp.1708–1713.
Shaefer, C. and Smith, S. (1990) The ARGOT Strategy II: Combinatorial Optimization, Thinking
Machinesnical Report RL90-1.
Shingo, S. (1989) A Study of the Toyota Production System, Productivity Press, New York.
Steinberg, J.G. and De Tomi, G. (2010) ‘Lean mining: principles for modelling and improving
processes of mineral value chains’, Int. J. of Logistics Systems and Management, Vol. 6,
No. 3, pp.279–298.
554 P.G. Saleeshya et al.

Welo, T., Tonning, O.R.B. and Rølvåg, T. (2013) ‘Lean systems engineering (LSE): hands-on
experiences in applying LSE to a student eco-car build project’, Conference on Systems
Engineering Research (CSER’13), Vol. 16, pp.492–501.
Whitley, D., Starkweather, T. and Fuquay, D. (1989) ‘Scheduling problems and traveling salesman:
the genetic edge recombination operator’, Proceedings of the Third International Conference
on Genetic Algorithms, pp.133–140.
Wisner, J. and Stanley, L. (2008) Process Management: Creating Value along the Supply Chain,
Thomson South-Western, Thomson Higher Education, 5191 Natorp Boulevard, Masson, OH
45040, USA.
Womack, J.P. and Jones, D.T. (2003) Lean Thinking: Banish Waste and Create Wealth in Your
Corporation, 2nd ed., Free Press, New York.
Zapp, M., Hoffmeister, M. and Verl, A. (2013) ‘Methodology to apply semantic wikis as lean
knowledge management systems on the shop floor’, 8th CIRP Conference on Intelligent
Computation in Manufacturing Engineering, CIRP 12, pp.444–449.