Evaluating Lean Performance of Indian Small and Medium Sized Enterprises in Automotive Sector

2.1 History

Toyota production system is the foundation of lean production system. The term lean was first put forth by MIT in 1990 to redefine Toyota production system as lean manufacturing system (Yadav et al., 2010 [6]).

Lean was the strategy to reduced cost, while continuously improving the quality of all kind of product and services you delivered, resulting in enhanced customer satisfaction (Womack & Jones, 2010 [7]).

The term leanness gives the measure of lean and also defined in variety of ways by the different researchers across the globe. Leanness was defined as effects of various lean tools and practices implemented on lean performance of the organizations (Vimal & Vinodh, 2012 [8]). Leanness was also defined as level of commitment to lean manufacturing (Soriano & Forrester, 2002 [9]). Leanness is a broader is broader in scope and transcend beyond the use of lean tools and practices to almost all operational features of the organization including design, development, quality, maintenance, and safety (Singh et al., 2010 [10]). Leanness defined as level of perfection in value stream mapping (Wan & Chen, 2008 [11]). Leanness was the measure of extent of waste elimination activities, continuous improvements, and achieving zero defect. Leanness defined as extent of waste elimination generating from redundant processes, inefficient operations, and excessive build-up of inventory (Hallgren & Olhager, 2009 [12]).

Lean philosophy was initially started as an operational stage in auto sector in 1980s. Toyota has become number one auto maker after successful implementation of lean philosophy, surpassing auto giant General motors, Ford motors, and Chrysler. Now lean production system started to gain acceptance worldwide in variety of industries including services, in quest of same benefits as Toyota enjoys (Piercy & Rich, 2009 [13]). It has been argued lean production system is more suitable for discrete manufacturing in comparison to continuous production found in process industry. But there are evidences of lean manufacturing system being successfully implemented in process industries like chemicals and pharmaceuticals with fair amount of success (Abdulmalek & Raj Gopal, 2007 [14]).

2.2 Identification of lean tools and practices

Commonly, the effects of five most vital lean practices were analysed in the literature such as, pull production system, heijunka, continuous improvement, total productive maintenance (TPM) and value stream mapping (VSM) on contemporary measures lean performances. Findings indicated that heijunka was having the strongest impact on operational performance, followed by continuous improvement, total productive maintenance, and value stream mapping. Research offered statistically proven evidence of linkage between lean practices and operational performance. Research would be helpful in better lean planning and lean tool selection (Belekoukias, 2014 [15]). Yadav and associates defined lean practices as total elimination of waste from the system, wherever it exists. Stream lined flow of information, material, and continuous improvement in all parts of the system were cited as key lean practices to achieve perfection (Yadav et al., 2010 [6]). Kaizen, cellular manufacturing, and poka-yoke were cited as key lean practices towards attaining the targeted level of leanness (Womack & Jones, 2010 [7]). According to a recent study conducted in 2014, lean root map was developed for describing the details of framework of lean implementation (Sundar et al., 2014 [16]). Continuous improvement was also suggested as one of the key lean practices. The goal of continuous improvement can be achieved by proper training, early detection of process problems, proper development of idea management, and adequate reward & recognition system (Holtskog, 2013 [17]). Various practices of lean implementation such as VSM, 5S, Poka-yoke, Kanban, JIT etc. were analysed for gear shaft manufacturing company. VSM was identified as key lean practices in variety of industrial environment (Nallusamy, 2016 [18]). Importance of takt time was emphasized through value stream mapping in order to synchronize the production with demand in order to achieve manufacturing superiority and dramatic improvements in the productivity of machine tool sector SMEs (Ricondo et al., 2016 [19]). Value stream mapping, 5S, and Kanban have been successfully applied to reduce lead time and in process inventory in Indian bearing industry, resulting in significant enhancement in the productivity (Saraswat et al., 2015 [20]). Value stream mapping combined with other lean tools such as 5 why and Ishikawa diagram were successfully applied in health care products to reduce lead time and in process inventory in the system (Tomas & Martin, 2015 [21]). Method time measurement & line balancing were successfully applied in truck body assembly line to reduce bottleneck and cycle time in the system (Kumar & Kumar, 2014 [22]). Value stream mapping, Kaizen, and single piece flow were applied to reduce cycle time, inventory and to increase visibility in crank shaft assembly line system (Venkataraman, 2014 [23]). Total productive maintenance was implemented in machine shop to enhance machine availability, performance of and to improve the quality yield (Singh et al., 2013 [24]). Lean practices such as waste elimination, kaizen, just in time, pull, cross functional team, decentralized responsibilities, integrated functions and vertical information systems were implemented in kitchenware industry to achieve considerable improvement in all contemporary performance parameters (Subashini & Kumar, 2014 [25]). Combination of lean practices and tools such as Total Productive Maintenance, Total Quality Management, Failure Mode and Effect Analysis, 5S, Quality Function Deployment, Kaizen, Kanban and Value Stream Mapping were used as inputs to structural equation modelling to determine various dimension of lean manufacturing (Vinodh & Joy, 2012 [26]). Simulation along with value stream mapping was used to design lean production system for doors and windows manufacturer following job shop production system (Gurumurthy & Kodali, 2011 [27]). Seventeen lean practices and lean tools namely: cross functional employee, quality circle, set up time reduction, 5S, Kanban, continuous flow, preventive maintenance, small lot size, total quality control (TQC), Kaizen, cellular layout, standard operational Training, focused factory, Supplier management, Visual control, Teamwork were suggested with varying degree of importance for successful lean implementation in SMEs environment (Rose et al., 2011 [28]). According to research conducted in Malaysia, Kanban system was the key lean practice for the successful lean implementation in Malaysian SMEs (Rahman et al., 2013 [29]). Value stream mapping, Kanban, 5S, Total productive maintenance, and single minute exchange of die, and Poka-yoke can be applied in Indian SMEs to reduce waste in the system and productivity improvements (Marasini et al., 2014 [30]). Questionnaire administered survey and subsequent statistical analysis suggested the most important and extensively used lean practices were continuous improvement, organization at work place, and standardization in Malaysian automotive SMEs (Rose et al., 2013 [31]). Exploratory factor analysis was also conducted for Malaysian manufacturers to identify the most important lean factors. Lean factors identified were process optimization, layout, and work organization (Yusup et al., 2016 [32]). A set of basic nine lean practices were suggested as starting point of the lean implementation affecting overall performance of the firm regardless of its size. These practices were focused on waste identification and elimination (Yusup et al., 2015 [33]). In Table 1 describes the objectives of various key lean tools.

Improper lean tool selection and lack of lean implementation methodology were cited as major hindrance in implementing lean philosophy in SMEs. It is quite evident from the literature survey; different lean practices were suggested and also feasible in different kinds of manufacturing environment but most of lean practices cited were overlapping and confusing in nature. Recently a comprehensive study was done in India to identify various lean factors and their associated lean variables feasible in Indian SMEs automotive industry environment. The following table (Table 2) describes various lean factors and associated lean variables for Indian SMEs in automotive sector, which provided the basis for comprehensive evaluation of leanness of Indian automotive SMEs (Tiwari & Tiwari, 2018 [34]).

2.3 Lean Measurement System

Slack based lean measurement system based on linear programming was developed with bench marking to quantify the degree of leanness, commonly known as lean score. Lean score determine the impact of various lean practices by quantifying the extent of value addition and waste elimination (Wan & Chen, 2008 [11]). Lean evaluation model was developed using fuzzy set theory, based on five vital lean parameters for Indian auto component industry (Singh et al., 2010 [10]). Fuzzy logic based theoretical lean measurement system was proposed comprises of three levels namely: lean enabler, lean criteria, and lean attribute followed by gap analysis. By identify the weak lean criteria which help to take corrective measures to improve the leanness organization (Vimal & Vinodh, 2012 [8]). Mathematical lean measurement model was developed to assess the effectiveness of various lean strategies on waste reduction in manufacturing organization. This model helped in identification of suitable lean strategy for manufacturing organization (Soriano & Forrester, 2002 [9]). Survey based method was used to evaluate the operational performance of variety of manufacturing firms spread in all four geographic regions in India, who have implemented lean philosophy. About 80% of the firms implemented the lean practices such as: customer focus, JIT, SMED, statistical process control, TPM, and cross functional problem solving. Implementation of these lean practices resulted in improvement of productivity and reduction in lead time, inventory & space requirement (Ghosh, 2012 [35]). Fuzzy logic based advisory system was formulated to help out SMEs in decision making with regards to the lean implementation. The system intended to measure the lean readiness of the organization. The level of lean readiness determined the amount of resources needed and having a vital bearing over the success and failure of lean implementation program (Achanga et al., 2012 [36]). Using attributes of both efficiency and effectiveness in manufacturing organization, lean evaluation system was developed to continuously assess lean implementation (Karim & Zaman, 2013 [37]). Lean model that takes into accounts the various lean principles was developed. The rationale behind the development of the model was in depth understanding of the change process in lean implementation leading to alleviation of uncertainty and vagueness in lean implementation. The model also provided the vital clues about the effect of inclusion and removal of a particular lean tool on the leanness and subsequently on operational performance of the company (Karlsson & Åhlström, 1996 [38]). Framework for evaluation of cost effectiveness of various lean tool was developed, applying multi criteria decision making. The framework was developed with the help of Deming & Kaizen cycle. Quality function deployment, value stream mapping and balanced score cord were applied in order to measure the cost effectiveness of lean tools using in relation to the customer requirements and the issues related to implementation (Alsyouf et al., 2011 [39]).

2.4 Lean measurement model

Customized multi-level lean measurement model using fuzzy comprehensive analysis was proposed to develop comprehensive lean measurement system for Indian automotive SMEs. (Li et al., 2009; Chang, 1996 [40, 41]). The basic assumption of model is that whole problem can be divided into three hierarchical level i.e. level 1 (Top level), level 2 & level 3. Assumption is also the limitation of the model allowing one-way communication that is top level to bottom level. Mean elements at each level of hierarchy particularly at lower level of hierarchy are independent in nature.

The following are step by step approach to model development: -

1. Establish target and identify the subject

   Target is to achieve sustainable strategic advantage to Indian SMEs in automotive sector by implementing lean philosophy.

2. Find evaluation grade to assess leanness.

   Evaluation grade set V = (Excellent v1, Very good v2, Good v3, Moderate v4, Poor v5)

3. Identification of various lean factors

   Next important step to develop lean measurement model was to define various lean factors those explain lean performance of Indian SMEs in automotive sector.

   F={F1,F2,F3,F4,F5,F6,F7}

   Where F = Lean performance.

4. Identification of various lean sub factors

   Lean sub factors are basically associated lean variables to factors.

   F1={f11,f12,f13,f14,f15,f16}F2={f21,f22,f23,f24,f25,f26}F3={f31,f32,f33,f34,f35}F4={f41,f42,f43,f44,f45}F5={f51,f52,f53,f54}F6={f61,f62,f63,f64}F7={f71,f72,f73}

5. Assign weight to various factors and associated sub factors applying fuzzy AHP (Refer section 2).

   AHP is extremely productive instrument in adjusting the fuzziness of the data. The linguistic level of evaluation by the experts are mapped to triangular fuzzy numbers (Table 4) to form fuzzy pair wise comparison matrix (Table 5). The model is having three level, level 1(Goal), level 2 (factors), & level 3 (Sub factors) (Refer table 5).

   Table 4

   Triangular Fuzzy conversion scale

   Linguistic Term	Membership function (l, m, n)
   Equally preferred	(1,1,1) (1,1,2)
   Moderately preferred	(1,2,3) (2,3,4)
   Strongly preferred	(3,4,5) (4,5,6)
   Very strongly preferred	(5,6,7) (6,7,8)
   Extremely preferred	(7,8,9)

   Table 5

   Pair wise comparison matrix

   	F1	F2	F3	F4	F5	F6	F7
   F1	[1   1   1]	456456456345567	112112112111112	123123123123112	123123112112123	111111111111112	111111/2111111/2111
   F2	1/41/51/61/41/51/61/41/51/61/31/41/51/51/61/7	[1   1   1]	1/21/31/31/21/31/41/21/31/41/31/41/51/31/41/5	11/21/311/21/31/21/31/411/21/31/21/31/4	1/21/31/41/21/31/411/21/31/21/31/41/21/31/4	1/51/61/71/51/61/71/31/41/51/31/41/51/31/41/5	1/51/61/71/51/61/71/51/61/71/61/71/81/61/71/8
   F3	111/2111/2111/2111111/2	234234234345345	[1   1   1]	123123123112112	112112123112111	11/21/3111111111/2111/2	11/21/311/21/311/21/3111/2111
   F4	11/21/311/21/311/21/311/21/3111/2	123123234123234	11/21/311/21/311/21/3111/2111	[1   1   1]	111111/2111/211/21/3111	11/21/311/21/311/21/311/21/3111/2	1/21/31/41/21/31/41/21/31/411/21/311/21/3
   F5	11/21/311/21/3111/2111/211/21/3	234234123234234	111/211/21/3111/2111/2111	111112112123111	[1   1   1]	11/21/311/21/311/21/311/21/3111	11/21/311/21/311/21/311/21/3111/2
   F6	111111111111111/2	567567345345345	123111111112112	123123123123112	123123123123111	[1   1   1]	111/2111/2111/2111/2111
   F7	111112111112111	567567567678678	123123123112111	234234234123123	123123123123112	111112112112111	[1   1   1]

6. Frame fuzzy membership matrix for each of the single factor

   Establish fuzzy membership matrix of each of the single factor on fn sub factors.

   Ri is the fuzzy evaluation matrix of factor Fi and r_ijk signifies degree of kth kind of evaluation for factor Fi to evaluation level j that is associated sub factor fij.

7. Frame fuzzy evaluation matrix (Judgment Vector) for each of the single factors

   Synthesize fuzzy evaluation matrix for every single factor using weights of associated sub factors (level 2) obtained with the help of fuzzy AHP that is wi = (w1, w2, w3, −−−−−−−, wn).

   Bi=(Wi×Ri)=(bi1,bi2,−−−−−−,bim)

   Fuzzy membership vector Bi represent mapping of comment set to factor Fi.

8. Perform fuzzy comprehensive analysis & obtain overall fuzzy membership matrix for the entire system.

   R=B1B2⋮⋮Bi=b11b12⋯⋯b1mb21b22⋯⋯b2m⋮⋮⋯⋮⋮⋮⋮⋯⋮⋮bn1bn2⋯⋮bnm

   Suppose the weights obtained for all the factors set using fuzzy AHP is W = (W1, W2, −−−−−−−, Wn).

   The overall fuzzy evaluation matrix or synthesize fuzzy matrix vector for the entire index system can be represented as: B = W × R = (b1, b2, −−−−−−, bm)

   B represents the mapping of comment set to entire index system. Suppose b₀ = max(b1, b2, −−−−−−, bm) corresponds to comment set v_k then v_k represented the evaluation of entire system.

3.1 AHP procedure

1. Frame pair wise comparison matrix using expert’s opinions.

   M11M12⋯⋯M1nM21M22⋯⋯M2n⋮⋮⋮⋮⋮⋮⋮⋮⋮⋮Mn1Mn2⋯⋯Mnn(1)

2. Consolidate the triangular fuzzy number with the help of following formula: -

   Mij=(Mij1⊗Mij2⊗−−−⊗M)1/n(2)

3. Perform consistency check for the decision making

   Consistency value CVn=CInRI≤0.1, where consistency index In=λmax−nn−1,λ max is the greatest Eigen value of pair-wise comparison matrix and n is the order of the matrix. RI stands for random index and its value can be found with the help of following table (Table 3). RI is the average value of CI. Table was obtained using the Satty scale developed by Thomas L. Satty applying simulation with the help of 100 runs.

4. calculate the consolidation grade using the following equation: -

   Ci=∑j=1mMij∑i=1n∑j=1mMij−1(3)

5. Compare the consolidated triangular fuzzy number and calculate the consolidation grade using the following equation:

   V(Ci≥Cj)=1ifmi≥mj0iflj≥ui〈lj−ui〉÷〈mi−ui〉−〈mj−lj〉Otherwise(4)

6. Calculate the minimum consolidation grade

   MinV〈C≥C1,C2⋯Ck〉MinV〈(C≥C1)and(C≥C2)⋯and(C≥Ck)〉Supposewi=MinV(Ci≥Cj)wherei≠k.(5)

   The weight vector can be framed as W = Normalize the weight vector to get final priority vector.

3.2 Triangular membership function

Triangular membership function was framed using fuzzy set theory. These membership functions were used for pair-wise comparison. Team comprises of five lean experts was chosen for the current study to rate their preferences using following membership function. Subjective evaluation of experts expressed in linguistic terms mapped to their corresponding fuzzy values (Table 4).

Mijk represent subjective importance of i^th criterion in comparison to j^th criterion by the k^th expert using membership function. The following equation computes the value of M_ij.

Where n is the number of experts involved in decision making.

4.1 Weight indexing

Weight assignment for F = (F₁, F₂, F₃, F₄, F₅, F₆, F₇), Where F = Lean performance using fuzzy AHP method. A panel of five lean experts was formed. Lean expert are consultants having hands on experience of at least four years of lean implementation in Indian SMEs.

1. Perform pair-wise comparison and translate the opinion of the experts into triangular fuzzy numbers (Table 5).

2. Perform consistency analysis of pare-wise comparison matrix.

   The maximum Eigen value of pair-wise comparison matrix is 7.36.

   CIn=λmax−nn−1,n=7CI(n)=0.06

   Value of RI = 1.32

   ConsistencyvalueCVn=CInRI⇒CV(n)=0.045

   Consistency value is less than 0.1, mean it can be safely assumed the decision taken by the expert in formulation of pair-wise comparison matrix is consistence in nature.

3. Using equation (3), consolidation of triangular fuzzy numbers gives the following sets of values:

   C2=(0.026,0.04,0.07),C1=(0.11,0.18,0.31),C3=(0.09,0.14,0.07),C4=(0.05,0.10,0.10),C5=(0.06,0.12,0.20),C6=(0.10,0.19,0.30),C7=(0.14,0.23,0.40)

4. Using equation (4, 5) calculate the consolidation grade.

   MinV(C1≥C2,C3,C4,C5,C6,C7)=0.77,MinV(C2≥C1,C3,C4,C5,C6,C7)=0.11,MinV(C3≥C1,C2,C4,C5,C6,C7)=0.47,MinV(C4≥C1,C2,C3,C5,C6,C7)=0.20,MinV(C5≥C1,C2,C3,C4,C6,C7)=0.35,MinV(C6≥C1,C2,C3,C4,C5,C7)=0.78,MinV(C7≥C1,C2,C3,C4,C5,C6)=1,

   Resulting weightvector = (0.77, 0.11, 0.47, 0.20, 0.35, 0.78, 1) and the normalization of weight vector gives the priority vector depicting the relative importance of the factors in relation to the degree of leanness. (F₁, F₂, F₃, F₄, F₅, F₆, F₇) = (0, 22, 0.03, 0.13, 0.054, 0.95, 0.21, 0.27). Similarly, the priority vectors for all the sub factors defined as level 3 factors can be calculated in relation to their factor at level 1 using the methodology as described above.

Using fuzzy AHP methodology, the weight index for the entire measure system was developed. The following table describes the weight index for the measurement system (Table 6).

4.2 Fuzzy comprehensive analysis

A panel of 50 experts from 50 Indian automotive SMEs is selected to evaluate the status of leanness SMEs at automotive sector in India and further to determine the way of implementation of various lean practices. All experts are managers in their industry and responsible for the key production line. Fuzzy evaluation matrix for factors F1 to F7 are framed according to evaluation grade described in the model development section.

For example, f₁₁ level 3 factor in concerned, 9 experts think it good, 16 experts think it moderate, and 25 experts think it poor. Fuzzy vector for f₁₁ written as, R₁₁ = (0.0, 0.0, 0.18, 0.32, 0.50). Similarly, fuzzy vectors for others are as follows:

Fuzzy evaluation matrix for factor F₁ is: -

Similarly, the fuzzy evaluation matrix for other factors can be written as:

Weight vectors for secondary factors (Level 3) are (Table 5):

Fuzzy vector for factor F₁ is B₁ = W₁ × R₁ = (0, 0.025, 0.155, 0.32, 0.50). Similarly, for other factors:

Fuzzy evaluation matrix for the entire measurement system can be framed as:

The evaluation grade is translated into numerical score with the help of table 7.

Evaluation grade V = {Excellent, Very good, Good, Moderate, Poor} = {95, 80, 65, 50, 20}

Weight vector for the entire measurement system is W = (0.22, 0.03, 0.13, 0.054, 0.095, 0.21, 0.27). Fuzzy vector for the entire measure system can be framed as:

B = W × R = (0.0, 0.06, 0.176, 0.31, 0.38). Maximum value of B is 0.38, which corresponds to poor (v5) in the comment set. Means the current status of leanness of Indian automotive industry is poor, which corresponds to numerical value of B × V = 39.34. Similarly, numerical score of evaluation of factors F1 to F7 are {38.075, 30.58, 43.78, 29.94, 38.3, 53.04, 33.82}.