Decision-making through Fuzzy TOPSIS and COPRAS approaches for

lean tools selection: A case study of automotive accessories

manufacturing industry

M Bhuvanesh Kumar*a, b, Jiju Antonyc, Elizabeth Cudneyd, Sandy L.

furterere, Jose Arturo Garza-Reyesf and S. M. Senthilb
a
Department of Production Engineering, National Institute of Technology
Tiruchirappalli, TN, India.
b
Department of Mechanical Engineering, Kongu Engineering College, TN, India.
c
Department of Industrial and Systems Engineering, Khalifa University, Abu Dhabi,
UAE.
d
John E. Simon School of Business, Maryville University, St. Louis, MO, USA.
e
Department of Engineering Management, Systems & Technology, University of
Dayton, Dayton, OH, USA.
f
College of Business, Las and Social Sciences, The University of Derby, Derby, UK.
*
Corresponding author

Email: [email protected]

Abstract

Similarity in prioritization of lean tools (LTs) by different frameworks on the

same problem is a point of contention. The goal of the present research is to
address LTs selection problem through two commonly used multi-criteria
decision making approaches, namely the technique for order preference by
similarity to ideal solution (TOPSIS) and complex proportional assessment
(COPRAS). A framework involving value stream mapping and plant layout
through TOPSIS and COPRAS approaches to find the best possible LTs for an
automotive accessories manufacturing plant is developed and assessed in this
research. The obtained similarity of rankings between TOPSIS and COPRAS is
71.42% and the difference is 28.58%. Based on the assessment, systematic layout
planning (SLP) is selected as the most suitable LT and its implementation is
elaborated in detail. Significant reductions were obtained in lead time (16.44%),
non-value added time (61.03%), transportation distances (40.42%), and waiting
time (86%). Additionally, lean implementation resulted in reduced inventory,
reduced internal traffic, improved productivity, and improved customer service.
 The LTs selection framework presented in this research work addresses the
computational complexity associated with the existing models and allows the
segregation of most preferable and least preferable criterion which eliminates the
criteria weight generation methods.

Keywords: value stream mapping; lean manufacturing; TOPSIS; COPRAS;

systematic layout planning; automobile industry.

1. Introduction

After World War II, the problems faced by Japanese manufacturers were different from

other western countries such as financial, massive shortages of material, and human

resources. The Toyota Production System (TPS) was introduced by Taiichi Ohno in the

1950s. In the late 1980s, the Lean Production System was developed from the TPS to

make it more suited to the need of western manufacturing companies (Womack, Jones,

& Roos, 1990). Seven basic wastes were identified in the TPS and recent studies have

explored additional wastes such as under-utilization of people and under-utilization of

facilities (Bhuvanesh Kumar & Parameshwaran, 2018). Most implementation practices

revealed value stream mapping (VSM) as the best tool to identify wastes from an

organization (Rohani & Zahraee, 2015). VSM has been practiced more frequently over

recent decades and it is very effective in reducing the lead time for manufacturing

organizations (Zahraee, Esrafilian, Kardan, Shiwakoti, & Stasinopoulos, 2021; Zahraee,

Tolooie, Abrishami, Shiwakoti, & Stasinopoulos, 2020). Adapting lean principles

greatly reduces non-value added (NVA) activities and yields cost savings in all

operations of the manufacturing environment (Alefari, Almanei, & Salonitis, 2020).

Practicing lean principles with appropriate training may create favorable impact

in the manufacturing industries in terms of their performance. There are many lean tools

(LTs) reported by researchers; however, organizations often have difficulty in choosing

the appropriate LT for implementation (Vinodh, Shivraman, & Viswesh, 2011).

Implementation of all the tools is a time consuming and costly process for SMEs; hence,

a systematic selection methodology is needed for the organizations (M Bhuvanesh

Kumar & Parameshwaran, 2019). To select the most appropriate LTs from the plethora

of tools, recent studies have proposed frameworks that combine industrial engineering

and optimization techniques (Devnath, Islam, Rashid, & Islam, 2020; M Bhuvanesh

Kumar & Parameshwaran, 2020; Rezaei, Rahiminezhad Galankashi, Mansoorzadeh, &

Mokhatab Rafiei, 2020). However, mathematical approaches used by these frameworks

require analytical skills, which are lacking in many SMEs. With respect to the efforts to

be made by the industrial practitioners, it is essential to assess different frameworks for

a single problem. The novelty of the present study lies in the assessment over different

frameworks for the LTs selection problem which is not available in the literature.

To address these needs, this research proposes a framework that employs VSM

and plant layout to identify and eliminate waste. Further, the LTs are prioritized and

assessed through two different multi criteria decision making (MCDM) approaches,

namely technique for order of preference by similarity to ideal solution (TOPSIS) and

complex proportional assessment of alternatives (COPRAS). To explore the ability of

LM practices in the organization, the research addresses the following objectives: (1) To

propose an LT selection framework using VSM, plant layout, TOPSIS, and COPRAS,

(2) To assess the sequence of LTs resulting from both fuzzy integrated TOPSIS and

COPRAS approaches, and (3) To corroborate the effectiveness of the proposed

framework by implementing it in the manufacturing industry.

By achieving the objectives through the application of proposed framework in a

case industry, the present work will contribute to the theory and practice of lean

implementation in SMEs. Compared to the similar frameworks on this topic, the present

research work minimizes the computational stages and more attention is given to the
 implementation practices. Also, the present study is substantiated through case study

which is not explored by many of the similar studies. The subsequent sections present

the literature review, methodology, and implementation through a case study. Finally,

the present and future state maps are compared to visualize the improvements, followed

by conclusion and future research directions.

2. Literature review

2.1 Review on lean implementation frameworks

The prioritization of LTs is considered as an MCDM problem and many articles have

demonstrated the application of integrated frameworks. The frameworks use MCDM

approaches such as fuzzy analytic hierarchy process (FAHP) (Susilawati, 2021), fuzzy

failure mode and effects analysis (FFMEA) (M Bhuvanesh Kumar & Parameshwaran,

2019), fuzzy TOPSIS (FTOPSIS) (Baskaran & Lakshmanan, 2019), fuzzy decision

making trial and evaluation laboratory (DEMATEL) (Seleem, Attia, Karam, & El-

Assal, 2020), VIseKriterijumska Optimizacija I Kompromisno Resenje (VIKOR) (Jing,

Niu, & Chang, 2019) and fuzzy COPRAS (FCOPRAS) (M. B. Kumar, Parameshwaran,

Antony, & Cudney, 2021), among others. Fuzzy enabled frameworks are arguably

better due to several advantages, while handling linguistic judgments with ambiguity

(Babaeinesami, Tohidi, & Seyedaliakbar, 2021). A large number of research articles use

triangular fuzzy numbers (TFNs) due to its computation simplicity (Susilawati, 2021;

Wang, 2021). The frameworks adapted by previous researchers are given in Table 1.

Table 1. Frameworks developed with respect to LTs selection and implementation

Research articles Integrated Major LTs/ Limitations/ disadvantages
frameworks solutions selected
Aouag, Soltani, and Fuzzy DEMATEL- Kaizen, Kanban, TPM  A small number of indicators were
Mouss (2021) FQFD considered.
M. B. Kumar et al. FAHP-COPRAS Plan-do-check-act (PDCA),  Dependent on the knowledge of
(2021) TPM, 5S, Kaizen expert team.
M Bhuvanesh Kumar FFMEA, AHP, and 5S, Kaizen, Layout  Not a generic application.
and Parameshwaran QFD modification  Cost-wise benefits are not
(2020) emphasized.
Devnath et al. (2020) QFD-TOPSIS Kanban, Cellular  Vagueness associated with expert
manufacturing, Kaizen opinion was not addressed
 No validation.
Seleem et al. (2020) Fuzzy DEMATEL Kaizen, 5S, Standardization  Impact of the approach was
subjective.
Prasad, Dhiyaneswari, VSM and root cause 5S, Kanban, Kaizen  Not a generic approach.
Jamaan, Mythreyan, analysis
and Sutharsan (2020)
Baskaran and FTOPSIS Just-in-time manufacturing,  Results were subject to change
Lakshmanan (2019) One piece flow with an increasing number of
participants.
Jing et al. (2019) Improved VIKOR Group technology, Mixed  Only one method was used to
method flow production evaluate the criteria.
Yadav, Seth, and Desai AHP-PROMETHEE LSS project tracking and  Increased complexity
(2018) review, Organization culture,  Elimination of risk factor.
Quality check systems
Belhadi, Touriki, and El AHP-TOPSIS Management participation,  Not included the interrelationship
fezazi (2017) Organizational culture, Key between barriers and solutions,
performance indicators (KPIs)

2.2 Choice of MCDM approaches

Most of the frameworks developed were two-phased and have a common attribute with

their methodologies. The first phase is to prioritize the wastes/barriers through relative

weights while the second phase prioritizes the solutions or LTs. Combined frameworks

integrating multiple MCDM techniques can create computational complications to

practitioners. Hence the methodology for the problem of LTs selection should be

simplified. The TOPSIS and COPRAS approaches classify attributes into two categories

such as most preferable/positive ideal solutions and least preferable/negative ideal

solutions (Varatharajulu, Duraiselvam, Kumar, Jayaprakash, & Baskar, 2021). Using

these tools, the relative weights can be directly applied which can eliminate the

requirement of the first phase from the two-phased computation procedure.

2.3 Gaps identified from the literature

The results of lean implementation frameworks for the same problem are questionable

when the similarity is brought into consideration. The result of one framework is not
compared with the other framework for the same problem. For example, the results of

FAHP-TOPSIS for a particular case organization may be compared with the result of

FQFD, FFMEA, PROMETHEE, or other combinations of these techniques. By doing

so, the interdependency between the wastes and LTs/solutions can be considered

generic and can give a better understanding. None of the papers have addressed this

concern. After a careful review of the literature, the research gaps identified are:

(1) Though TOPSIS and COPRAS approaches can simplify the LTs selection

problem as a single-phased approach, no research work has adapted these

methods independently to reduce computational complexity.

(2) None of the published research works have compared the results of different

MCDM approaches to the same case organization.

These research gaps created a scope for a novel approach to evaluate TOPSIS

and COPRAS approaches for the prioritization of LTs to the same problem.

3. Methodology

The three-phase methodology adapted in the present study is shown in Figure 1.

Recognition of wastes and LTs is carried out in phase I. Assessment of the MCDM

approaches is carried out in phase II. The implementation of LTs and the comparison of

current and future states are performed in phase III.

3.1 Phase I

Phase-I is associated with the development of the current state VSM. To better observe

the different forms of wastes, the plant layout is drawn along with the VSM. An expert

team consisting of a plant head and two industrial engineering professionals are selected

for the team to appraise the existing state appropriately. Upon conducting several
brainstorming sessions and observations, the wastes are identified. Suitable LTs are

selected from the literature and verified with industrial engineering professionals.

Existing state plant

layout and VSM
Production
environment
Wastes identification Phase I

Literature &
Selection of lean
expert
tools/techniques opinion

Prioritize lean tools Prioritize lean tools

using Fuzzy TOPSIS using Fuzzy COPRAS
Phase II

Assessment over the MCDM results

Lean implementation
Phase III
Future state plant layout and VSM

Figure 1. Methodology adapted

The experts are asked to input their opinion in a decision matrix to show the

significance level of a particular LT to eliminate a particular waste. In order to avoid

vagueness present in linguistic opinions, fuzzy logic with TFNs shown in Table 2 is

used (M. B. Kumar et al., 2021). The decision matrix from each expert are averaged to

form an all-inclusive synthetic dependency matrix ‘D’, as shown in Eq (1).

d d ⋯ d
d d d
D= (1)
⋮ ⋮ ⋱ ⋮
d d … d

Where, ‘ n’ is the number of identified wastes and ‘ m’ is the number of

identified LTs. Hence, ‘ d ’ is the combined dependency value among the LT, ‘ m’,

and waste, ‘ n’. Further, these TFNs are defuzzified into a crisp numbers using the best

non-fuzzy performance (BNP) using Eq (2) (M. B. Kumar et al., 2021).

 BNP = U −L + M −L ⁄3 + L (2)

Where, ‘ A ’ is the averaged value of TFN for the i waste. L , M , and U

are the lower, middle, and upper values of averaged TFN for the i waste. The newly

formed dependency/decision matrix is expressed as ‘ D ’ in the subsequent sections.

Table 2. The scale to show linguistic opinions along with equivalent TFNs
Linguistic opinion Equivalent TFN Reciprocal of TFN
Enormously high significance (EHS) (8, 9, 9) (1/9, 1/9, 1/8)
Very high significance (VHS) (7, 8, 9) (1/9, 1/8, 1/7)
High significance (HS) (6, 7, 8) (1/8, 1/7, 1/6)
Fairly high significance (FHS) (5, 6, 7) (1/7, 1/6, 1/5)
Fair significance (FS) (4, 5, 6) (1/6, 1/5, 1/4)
Fairly low significance (FLS) (3, 4, 5) (1/5, 1/4, 1/3)
Low significance (LS) (2, 3, 4) (1/4, 1/3, 1/2)
Very low significance (VLS) (1, 2, 3) (1/3, 1/2, 1)
Enormously low significance (ELS) (1, 1, 1) (1, 1, 1)

3.2 Phase II

3.2.1. TOPSIS

In the TOPSIS method, the agreeable result can be observed as preferring the response

with the closest distance from the positive ideal boundary and longest distance from the

adverse ideal boundary (Ghosh, Mandal, & Ray, 2021). The TOPSIS methodology uses

the following steps.

Step 1: The first step is the identification of dependent and independent

attributes (wastes). The dependent attributes that require maximization function are

referred to as the most preferable attributes. The independent attributes are those

required by the minimization function and referred to as the least preferable attributes.

Step 2: The information related to the attributes is expressed in the form of a

matrix, which is often referred to as the dependency matrix D = x ×

, which has i

rows (m - LT) and j columns (n - attributes). The decision matrix, ‘ D ’,for the present

work is given in Eq (3).

 x x … x
x x … x
D = … … … … (3)
x x … x

Step 3: The elements of the decision matrix are normalized using Eq (4). The

normalized matrix is expressed as ‘ N ’.

x
N = j = 1, 2, … , n
∑ (4)
x

Step 4: A weighted normalized decision matrix is determined by multiplying the

elements of the normalized decision matrix by its corresponding weights (W ) obtained

from the experts as shown in Eq (5). The elements of the resulted matrix forms a

weighted normalized matrix, ‘ W ’, using Eq (6) in which x ’ is the weighted elements..

W = N × W (5)

x x … x
x x … x
W = (6)
… … … …
x x … x

Step 5: The positive ideal solutions, A**, and negative ideal, A*, solutions are

calculated using Eqs (7) and (8).

A∗∗ = maxW jϵJ , minW jϵJ′ ; i = 1, 2, … , m ; J = 1, 2, … , n ; (7)

A∗ = minW jϵJ , maxW jϵJ′ ; i = 1, 2, … , m; J′ = 1, 2, … , n (8)

Where, ‘ J ’ is related to the most preferable attributes and ‘ J' ’ is related to the

least preferable attributes.

Step 6: The separation measure is determined for each LT from the most

preferable ideal solution using Eq (9). In the same way, the least preferable ideal

solution is given by Eq (10).

 S ∗∗ = W − A∗∗ where i = 1, 2, … , m (9)
,

S∗ = W − A∗ where i = 1, 2, … , m (10)
,

Step 7: Further, the relative distance is estimated to be the ideal solution using

Eq (11). A higher Ci* value represents better performance of the attributes

corresponding to the LTs.

S∗
C∗ = (11)
S ∗∗ + S∗

Step 8: Finally, ranking of the LTs is made based on the relative closeness value.

3.2.2. COPRAS

On the other hand, the COPRAS approach is also applied to prioritize the identified LTs

in association with the identified wastes. The detailed computation procedure of

COPRAS approach follows a five step procedure.

Step1: Constructing a defuzzified dependency matrix between m - LTs and n -

attributes (wastes) as shown in Eq (12).

d d ⋯ d
d d d
D= (12)
⋮ ⋮ ⋱ ⋮
d d … d

Step2: The elements of ‘ D ’ are normalized using Eq (13) to get a normalized

dependency matrix ‘ D ’ in which ‘ d ’ is an averaged element of ‘ D ’,. The resulting

normalized matrix is represented by Eq (14).

= ; =1 ; =1 n; (13)
∑
 d d ⋯ d
⎡ ⎤
D = ⎢⎢ d d d ⎥
⎥ (14)
⎢ ⋮ ⋮ ⋱ ⋮ ⎥
⎣d d … d ⎦

Step 3: The elements are multiplied by relative weights (w ) obtained from the

expert members to form a weighted normalized dependency matrix, ‘ D ’, as shown in

Eq (15). The resultant matrix is in the form of Eq (16).

d = d × w ; i = 1 to n, and j = 1 to m (15)

⎡d d ⋯ d ⎤
⎢ d ⎥
D = ⎢d d
⎥ (16)
⎢ ⋮ ⋮ ⋱ ⋮ ⎥
⎣d d … d ⎦

Step 4: The attributes required to be maximized are the most preferable and

given higher weights. The attributes required to be minimized are the less preferable and

are given lesser weights. Using Eqs (17) and (18), the summation of normalized values

of more preferable attributes, ‘ P ’, and least preferable attributes, ‘ R ’, are computed.

P = d (17)

Here, ‘ k ’ is the number of more preferable attributes. The number of less

preferable attributes can be derived as ‘ (n − k ) ’.

R = d (18)

Step 5: The final weights of the LTs are computed using Eq (19).

R ∑ R
Q = P+ (19)
R ∑

Step 6: Efficacy degree of each LT denoted as ‘ N ’ is calculated using Eq (20).

Q (20)
N = × 100%
Q
3.2.3. Assessment of MCDM approaches

The prioritized lists of LTs from both the TOPSIS and COPRAS approaches are

compared for their percentage of similarity of sequence.

3.3 Phase III

This is the implementation and realization phase where the selected tools from the top

order of prioritized list are implemented in the organization. According to the changes

made in the existing state, the future state plant layout and VSM are developed.

4. Case Study

The case study was conducted in automotive accessories manufacturing company

located in the southern part of India. They manufacture bumpers, footsteps, and luggage

carriers for the Indian automobile industry. The product quality and demand from the

Indian market enables them to stay at the top of the market.

4.1 Phase I: Development of current state VSM and plant layout

During the initial visits, direct observations and discussions with people from the case

organization were performed. The organization’s shop floor lacked an efficient process

flow due to unnecessary transportation and movement. Also, unplanned continuous

production irrespective of demand leads to overproduction of products that may become

obsolete. The current state plant layout is shown in Figure 2, and the current state VSM

is shown in Figure 3. The sequence of operations to make a steel car bumper are,

bending process, plate cutting operation, plate drilling process, plate leveling operation,

welding process, grinding process, 7-tank process, polishing and buffing processes, sub-

assembly of components, and packing. The material flow distance, layout details, and

operation times are noted by interaction with the manager, supervisors, and staff. A
close inspection is made to gather the data such as operating time (OT), changeover

(C/O), available time (AT), cycle time (CT), and number of operators. Following the

data collection, the plant layout was measured for its total area, relative position of

resources, and distances between different functional cells.

Figure 2. Existing plant layout

The VSM of the current state shown in Figure 3 illustrates the flow of materials

transported to successive operations, mode of communication from the manager, and

quantitative information related to each process. Various measures such as lead time

and NVA time are derived. Based on a complete study made from the plant layout,
current state VSM, and observations, the wastes identified are: excess inventory (W1),

unnecessary transportation (W2), material wastes (W3), underutilization of people

(W41), unnecessary motion (W5), waiting (W6), and long lead time (W7). Also, during

the transportation of work-in-process (WIP), internal traffic was high. With the help of

the experts’ opinion and the literature, the LTs identified for the case organization are

cross functional training (LT1), layout planning (LT2), one-piece flow (LT3),

standardization (LT4), 5S (LT5), kaizen (LT6), and workforce monitoring (LT7).

Figure 3. Current state VSM

4.2 Phase II: TOPSIS and COPRAS methods

The averaged experts’ opinions on the dependency between the wastes and LTs are

defuzzified using Eq (2). The obtained BNP values drawn from Eq (2) forms a synthetic

dependency matrix of 7 × 7 as shown in Table 3.

Table 3. Dependency matrix between wastes and LTs
LP/W W1 W2 W3 W4 W5 W6 W7
LT1 1.333 2.333 1.667 1.000 4.000 3.333 1.667
LT2 6.000 5.667 5.000 2.000 6.333 8.222 7.333
LT3 1.333 2.333 8.222 8.222 5.667 1.000 3.000
LT4 2.333 8.444 1.667 1.333 6.667 6.333 5.667
LT5 5.333 4.667 5.333 1.000 3.333 5.333 8.667
LT6 7.667 5.667 6.667 1.333 1.667 3.667 2.333
LT7 8.222 5.000 3.667 1.000 1.667 4.667 1.000

4.2.1Ranking of LTs using TOPSIS

The TOPSIS approach is further applied to rank the identified LTs.

Step 1: Waste reduction is the most preferable attribute; therefore, it requires

maximization function. Underutilization of people is the least preferable and needs the

minimization function. With experts’ help, the most preferable attributes are given with

a relative weight of 0.15 and least preferable attributes are given with a relative weight

of 0.1; therefore, the cumulative weight for all of the attributes is 1.

Step 2 : The dependency matrix of the present work is expressed as Eq (21).

1.333 2.333 … 1.667

[ × ] = 6.000 5.667 . 7.333
(21)
⋮ ⋮ . ⋮
8.222 5.000 … 1.000
Step 3: The dependency matrix is then normalized using Eq (4) to obtain the

normalized dependency matrix as shown in Eq (22).

0.094 0.167 … 0.124

[ × ] = 0.424 0.407 . 0.547 (22)
⋮ ⋮ . ⋮
0.581 0.359 … 0.074
Step 4: Based on the relative weights assigned to the attributes, a weighted

normalized decision matrix using Eq (5) is developed as shown in Eq (23).

1.333 2.333 … 1.667

6.000 5.667 . 7.333
× = (23)
⋮ ⋮ . ⋮
8.222 5.000 … 1.000
 Step 5: Using Eqs (7) and (8), the positive ideal solutions, A** and negative

ideal solutions, A*, are computed as shown in Eq (24).

∗∗
0.087 0.091 0.091 0.093 0.082 0.091 0.097
∗
= (24)
0.014 0.025 0.018 0.011 0.020 0.011 0.011

Step 6 & 7: Similarly, the separations measures and the relative distances are

calculated using Eqs (9)-(11) as shown in Table 4. Finally, the ranking of LTs is

performed based on the highest value of Ci*.

4.2.2 Ranking of LTs using COPRAS

The ranking of the LTs using the COPRAS approach is made by the subsequent steps.

Step 1: The decision matrix given in Eq (21) is drawn here for computations.

Step 2: The framed decision matrix is further normalized using Eq (13). The

resulted matrix is shown in Eq (25).

0.041 0.68
… 0.056
D = 0.186 0.166 . 0.247 (25)
⋮ ⋮ . ⋮
0.255 0.147 … 0.034
Step 3: Similar to the TOPSIS approach, underutilization of people is given the

least preference with a relative weight of 0.1 and other wastes are given the most

preferences with a relative weight of 0.15. The weighted normalized decision matrix

using Eq (15) is computed and presented in Eq (26).

0.006 0.010 … 0.008

D = 0.028 0.028 . 0.037 (26)
⋮ ⋮ . ⋮
0.038 0.022 … 0.005
Step 4: The summation of the most preferable, Pi, and least preferable, Ri,

normalized values are calculated using Eqs (17) and (18).

0.068 0.183 0.103 0.148 0.156 0.129 0.112

= (27)
0.006 0.013 0.052 0.008 0.006 0.008 0.006
 Steps 5 & 6: Finally, the weight of each LT is calculated and ranked based on

the efficacy degree of each project using Eqs (19) and (20) as presented in Table 4.

Table 4. Ranking of LTs using TOPSIS and COPRAS

LT TOPSIS Pi Ri Qi Ni COPRAS
S** S* C*
ranking ranking
Cross functional training (LT1) 0.160 0.039 0.194 7 0.068 0.006 0.072 38% 7
Layout planning (LT2) 0.088 0.122 0.582 1 0.183 0.013 0.190 100% 1
One piece flow (LT3) 0.127 0.120 0.484 2 0.103 0.052 0.130 68% 5
Standardization (LT4) 0.125 0.108 0.464 3 0.148 0.008 0.152 80% 3
5S (LT5) 0.114 0.082 0.420 5 0.156 0.006 0.159 84% 2
Kaizen (LT6) 0.117 0.099 0.458 4 0.129 0.008 0.133 70% 4
Workforce monitoring (LT7) 0.126 0.091 0.420 6 0.112 0.006 0.116 61% 6

4.2.3 Assessment over MCDM approaches

With reference to the computed results, the ranking of LTs obtained from the TOPSIS

methodology is LT2 > LT3 > LT4 > LT6 > LT5 > LT7 > LT1 and LT2 > LT5 > LT4 >

LT6 > LT3 > LT7 > LT1 by COPRAS. Layout planning (LT2) is ranked first with a

relative closeness value of 0.58 and cross functional training (LT1) is ranked lowest

with a relative closeness value of 0.19 in TOPSIS. Similarly, based on the COPRAS

calculations, layout planning (LT2) is ranked first with an efficacy degree of 100%, and

cross functional training (LT1) is ranked lowest with an efficacy degree of 38%. The

ranking of LTs using COPRAS has a greater occurrence with TOPSIS calculations. The

similarity between the rankings is 71.42% between both the MCDM approaches with

the same input. The remaining 28.58% in ranking differentiate the approaches.

4.3 Phase III

4.3.1. Lean implementation

The first ranked LT, layout planning (LT2), is initially selected to implement within the

organization. In order to modify the existing plant layout systematically, the SLP
method is adopted as it is a proper and structured planning method for designing an

effective plant layout (Ali Naqvi, Fahad, Atir, Zubair, & Shehzad, 2016). SLP is a tool

used to position a workstation in a production floor by locating the two regions that are

very close to one another with logical relationship and high frequency. For the present

case study, the data table is created as an initial step to collect the information related to

the distance in meters between the 12 consecutive operations as shown in Table 5. SLP

uses the relationship chart (REL) in which the diamond shaped cells carry the values for

the degree of closeness between workstations. A scale shown in Table 6 is used for this

purpose (Qamar, Meanazel, Alalawin, & Almomani, 2020). The pair-wise ratings of

closeness are inter-departmental and are used to compose an appropriate layout.

Table 5. Distance between the operations of workstations

Operations Distance between operations (ft)

1 2 3 4 5 6 7 8 9 10 11 12
Cutting process 4.0
Bending process 45.2
Plate cutting process 5.0
Drilling process 45.2
Leveling process 4.5
Welding process 28.0
Grinding process 5.0
7 Tank process 18.0
Polishing process 3.0
Buffing process 12.0
Sub Assembling 4.0
Packing process

Table 6. Alphabetical codes used for REL chart

S.No Alphabetical codes Description

1 A Absolutely necessary
2 E Especially important
3 I Important
4 O Ordinary
5 U Unimportant
6 X Undesirable
As part of the SLP procedure, a REL chart is developed with the help of number

of trips times and the distance between locations as presented in Figure 4. According to

the frequency and significance of the processes, the alphabetical codes from Table 6 are
used. From the activity relationship, it is clear that the pairs within the existing sequence

of operations has a strong relationship such as bending & plate cutting (A), drilling &

leveling (A), welding & grinding (A), and 7-tank process & polishing (E). The pairs of

other processes have a lesser important relationship (O). Shifting of departments within

these pairs without disturbing the production sequence can be done. Hence, a number of

alternative layouts were proposed by reallocating the processes based on the importance

from most to least. Among the proposed layouts, the best layout is chosen considering

the practical limitations.

Figure 4. Activity relationship chart

4.3.2 Modified layout & future state VSM

A modified layout based on the selection from the SLP is constructed as shown in

Figure 5. The major modification made on the existing plant layout is the relocation of

the plasma cutting operation to unused space found near the stock storage. Therefore,

the distance and transportation time from the welding booth were reduced to half.

Relocation of plasma cutting operation left an empty space in which the polishing,

buffing, and packing processes are accommodated. This enabled the packing process to

complete faster compared to the former layout. Inventories and stock yards were also

modified in the future state plant layout. The modified layout also leaves several empty

spaces that can support the future expansion plans. The future state VSM is drawn
according to the modified layout as shown in Figure 6. The modifications are made as

per the analysis conducted on the new state. The values of CT, C/O time, and AT are

updated in the future state VSM. The reduction of lead time and NVA time is observed

in the future state. The uptime (UT) for the processes is calculated using Eq (28).

AT − ( /O) (28)
UT =
AT

Figure 5. Modified plant layout

5. Sensitivity Analysis

The sensitivity analysis is carried out to compare and validate the MCDM approaches. It
will also address the associated uncertainty issues (Hasheminezhad, Hadadi, &

Shirmohammadi, 2021). It is conducted based on 3 cases as shown in Table 7 and

Figure 7. In order to represent different cases, the weights for experts’ opinion were

changed. For instance, weight for first expert is given 0.4 compared to other two experts

which is 0.3 for case 1. According to sensitivity analysis, the ranking remains

unchanged for both the MCDM approaches for all the cases expect the case 1 of

TOPSIS approach in which the ranking of LT5 and LT7 alone varied. Despite of an

insignificant variation, the rankings from MCDM approaches were stable and reliable.

Figure 6. Future state VSM

Table 7. The cases for sensitivity analysis of proposed framework

Criteria Expert 1 Expert 2 Expert 3

Original Weights 0.333 0.333 0.333
Case 1 0.4 0.3 0.3
Case 2 0.3 0.4 0.3
Case 3 0.3 0.3 0.4
 TOPSIS COPRAS
(a) (b)
7 7
6 6
Rank 5 5

Rank
4 4
3 3
2 2
1 1
CASE 1 CASE 2
0 0
CASE 3 Origina l ra nk
LT1 LT2 LT1 LT2
LT3 LT4 LT3 LT4
LT5 LT6 LT5 LT6
LT7 LT7

Figure 7. Sensitivity analysis of (a) TOPSIS approach, and (b) COPRAS approach

6. Discussion

The assessment among the rankings from TOPSIS and COPRAS approaches revealed

71.42% concurrency and 28.57% non-concurrency in the results. Most of the LTs are

ranked similar except LT3 (one piece flow) and LT5 (5S). This show the ranking by

these two methods is reliable; hence the top ranked LT (SLP) from both the approaches

was selected as a generalized solution and implemented in the case industry. A few

operations were relocated based on the importance in the relationship acquired from the

REL chart. The results after the lean implementation have been recorded in VSM of

future state. Comparison of the present and future state VSM showed great reductions in

lead time from 70.35 min to 58.78 min (16.44%), NVA time from 18.63 min to 7.26

min (61.03%), total transportation distance from 188 m to 112 m (40.42%), and WIP

waiting time from 646 s to 84 s (86%). Additionally, the enhanced layout reduced

internal traffic and inventories considerably. This will also improve employee

participation, efficiency and customer service.

Comparing with the similar studies in the topic, the framework developed in the

present study minimized the computation stages and the ranking results were more

reliable as validated through sensitivity analysis. Using the outcomes of this research,

the practicing managers can understand that ineffective allocation of resources lead to
the generation of wastes which can hinder the organizational performance. More

importantly, project managers should realize that systematic layout planning during the

design stage of production floor will reduce reallocation cost and minimize most of the

operation oriented wastes. Apart from the theoretical and practical contribution, this

study will also help to develop more number of lean implementation frameworks in

future.

7. Conclusion

This research work applied two different MCDM approaches, namely TOPSIS and

COPRAS, for a LT selection problem for an automotive accessories manufacturing

organization. The ranking results had more concurrency among them. After

implementing the selected LT from the ranking, the analysis was carried out by

comparing the current and future state VSMs to calculate definite improvements. Future

state map showed a significant reduction in lead time, NVA time, and transportation

time & distance. There were also some tangible benefits such as reduced internal traffic,

stagnation of parts, inventory, improved supervision, and saving of floor space obtained.

Besides the favorable outcomes, a few more suggestions are given as, (a) addition of

material handling equipment, (b) improve the packing section by providing tables, and

(c) replace the obsolete machines with new machines. The effectiveness of the

framework depends on the knowledge and experience of the experts; therefore, it is

subject to change. Future research should compare the results of other MCDM

approaches for different industries to confirm the generalization. Future studies can

include simulation and cost benefit analysis to validate the results.

Declaration of interest

The authors declare that there is no competing interest to disclose.

References
Alefari, M., Almanei, M., & Salonitis, K. (2020). Lean manufacturing, leadership and
employees: the case of UAE SME manufacturing companies. Production &
Manufacturing Research, 8(1), 222-243. doi: 10.1080/21693277.2020.1781704
Ali Naqvi, S. A., Fahad, M., Atir, M., Zubair, M., & Shehzad, M. M. (2016).
Productivity improvement of a manufacturing facility using systematic layout
planning. Cogent Engineering, 3(1), 1207296
Aouag, H., Soltani, M., & Mouss, M. D. (2021). Enhancement of value stream mapping
application process through using fuzzy DEMATEL and fuzzy QFD approaches:
a case study considering economic and environmental perspectives. Journal of
Modelling in Management, 16(3), 1002-1023. doi: 10.1108/JM2-01-2020-0007
Babaeinesami, A., Tohidi, H., & Seyedaliakbar, S. M. (2021). Designing a data-driven
leagile sustainable closed-loop supply chain network. International Journal of
Management Science and Engineering Management, 16(1), 14-26. doi:
10.1080/17509653.2020.1811794
Baskaran, S. M., & Lakshmanan, A. (2019). A framework model for lean tools selection
for improving material flow using fuzzy TOPSIS. International Journal of
Productivity and Quality Management, 27(2), 196-228
Belhadi, A., Touriki, F. E., & El fezazi, S. (2017). Prioritizing the solutions of lean
implementation in SMEs to overcome its barriers. Journal of Manufacturing
Technology Management, 28(8), 1115-1139. doi: 10.1108/JMTM-04-2017-0066
Bhuvanesh Kumar, M., & Parameshwaran, R. (2018). Fuzzy integrated QFD, FMEA
framework for the selection of lean tools in a manufacturing organisation.
Production Planning & Control, 29(5), 403-417. doi:
10.1080/09537287.2018.1434253
Devnath, A., Islam, M. S., Rashid, S., & Islam, E. (2020). An integrated QFD-TOPSIS
method for prioritization of major lean tools: a case study. International Journal
of Research in Industrial Engineering, 9(1), 65-76
Ghosh, S., Mandal, M. C., & Ray, A. (2021). Green supply chain management
framework for supplier selection: an integrated multi-criteria decision-making
approach. International Journal of Management Science and Engineering
Management, 1-15. doi: 10.1080/17509653.2021.1997661
Hasheminezhad, A., Hadadi, F., & Shirmohammadi, H. (2021). Investigation and
prioritization of risk factors in the collision of two passenger trains based on
fuzzy COPRAS and fuzzy DEMATEL methods. Soft Computing, 25(6), 4677-
4697
Jing, S., Niu, Z., & Chang, P.-C. (2019). The application of VIKOR for the tool
selection in lean management. Journal of Intelligent Manufacturing, 30(8),
2901-2912
Kumar, M. B., & Parameshwaran, R. (2019). Fuzzy weighted geometric mean
approach-based FMEA to prioritise lean failure modes in manufacturing
industries. International Journal of Manufacturing Technology and
Management, 33(6), 398-427
Kumar, M. B., & Parameshwaran, R. (2020). A comprehensive model to prioritise lean
tools for manufacturing industries: a fuzzy FMEA, AHP and QFD-based
approach. International Journal of Services and Operations Management, 37(2),
170-196. doi: 10.1504/ijsom.2020.110337
Kumar, M. B., Parameshwaran, R., Antony, J., & Cudney, E. (2021). Framework for
Lean Implementation Through Fuzzy AHP-COPRAS Integrated Approach.
 IEEE Transactions on Engineering Management, 1-13. doi:
10.1109/TEM.2021.3089691
Prasad, M. M., Dhiyaneswari, J., Jamaan, J. R., Mythreyan, S., & Sutharsan, S. (2020).
A framework for lean manufacturing implementation in Indian textile industry.
Materials Today: Proceedings
Qamar, A. M., Meanazel, O. T., Alalawin, A. H., & Almomani, H. A. (2020).
Optimization of Plant Layout in Jordan Light Vehicle Manufacturing Company.
Journal of The Institution of Engineers (India): Series C, 101(4), 721-728. doi:
10.1007/s40032-020-00576-5
Rezaei, A., Rahiminezhad Galankashi, M., Mansoorzadeh, S., & Mokhatab Rafiei, F.
(2020). Supplier Selection and Order Allocation with Lean Manufacturing
Criteria: An Integrated MCDM and Bi-objective Modelling Approach.
Engineering Management Journal, 32(4), 253-271
Rohani, J. M., & Zahraee, S. M. (2015). Production Line Analysis via Value Stream
Mapping: A Lean Manufacturing Process of Color Industry. Procedia
Manufacturing, 2, 6-10. doi: https://doi.org/10.1016/j.promfg.2015.07.002
Seleem, S. N., Attia, E.-A., Karam, A., & El-Assal, A. (2020). A lean manufacturing
road map using fuzzy-DEMATEL with case-based analysis. International
Journal of Lean Six Sigma, 11(5), 903-928. doi: 10.1108/IJLSS-12-2017-0147
Susilawati, A. (2021). Productivity enhancement: lean manufacturing performance
measurement based multiple indicators of decision making. Production
Engineering, 15(3), 343-359
Varatharajulu, M., Duraiselvam, M., Kumar, M. B., Jayaprakash, G., & Baskar, N.
(2021). Multi criteria decision making through TOPSIS and COPRAS on
drilling parameters of magnesium AZ91. Journal of Magnesium and Alloys. doi:
https://doi.org/10.1016/j.jma.2021.05.006
Vinodh, S., Shivraman, K., & Viswesh, S. (2011). AHP-based lean concept selection in
a manufacturing organization. Journal of Manufacturing Technology
Management, 23(1), 124-136
Wang, F. (2021). Preference degree of triangular fuzzy numbers and its application to
multi-attribute group decision making. Expert Systems with Applications, 178,
114982. doi: https://doi.org/10.1016/j.eswa.2021.114982
Womack, J. P., Jones, D. T., & Roos, D. (1990). Machine that changed the world:
Simon and Schuster.
Yadav, G., Seth, D., & Desai, T. N. (2018). Application of hybrid framework to
facilitate lean six sigma implementation: a manufacturing company case
experience. Production Planning & Control, 29(3), 185-201. doi:
10.1080/09537287.2017.1402134
Zahraee, S. M., Esrafilian, R., Kardan, R., Shiwakoti, N., & Stasinopoulos, P. (2021).
Lean construction analysis of concrete pouring process using value stream
mapping and Arena based simulation model. Materials Today: Proceedings, 42,
1279-1286. doi: https://doi.org/10.1016/j.matpr.2020.12.955
Zahraee, S. M., Tolooie, A., Abrishami, S. J., Shiwakoti, N., & Stasinopoulos, P.
(2020). Lean manufacturing analysis of a Heater industry based on value stream
mapping and computer simulation. Procedia Manufacturing, 51, 1379-1386.
doi: https://doi.org/10.1016/j.promfg.2020.10.192