Standardizing Components and Rotating Workers Using GT-Based Algorithm—A Case...

[Full title: Standardizing Components and Rotating Workers Using GT-Based Algorithm—A Case Study]

sustainability Article Standardizing Components and Rotating Workers Using GT-Based Algorithm—A Case Study Ming-Liang Li Citation: Li, M.-L. Standardizing Components and Rotating Workers Using GT-Based Algorithm—A Case Study Sustainability 2021 , 13 , 7880 https://doi.org/10.3390/su 13147880 Academic Editors: Mitsutaka Matsumoto and Kenichiro Chinen Received: 11 June 2021 Accepted: 8 July 2021 Published: 14 July 2021 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations Copyright: © 2021 by the author Licensee MDPI, Basel, Switzerland This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/) Department of Industrial Management, Oriental Institute of Technology, 58, Sec.2, Sihchuan Rd., Banqiao Dist., New Taipei City 22061, Taiwan; fg 017@mail.oit.edu.tw; Tel.: +886-(2)-7738-0145 (ext. 5113) Abstract: In both manufacturing and remanufacturing systems, exploiting bulk buying and avoiding delivery delays due to material shortages are crucial. One method that aids in these processes is component standardization. Additionally, company managers seek to reduce labor costs and mitigate the risk of sudden worker resignation or absence due to, for example, reasons associated with the COVID-19 pandemic. The aforementioned problems could be solved using the sorting algorithm proposed in this study. The concept of the proposed algorithm is based on group technology. One numerical example and two case studies are presented to demonstrate the utility of the proposed algorithm. The first example suggested that the performance of the algorithm proposed in this study is superior to another one in the literature. The second one demonstrated that the algorithm in this work achieves component standardization by reducing an initial number of 12 components down to 6. The final case study provides an effective means of grouping workers with similar operational abilities and suggests how to assign new tasks to other skilled workers if a worker resigns suddenly or cannot attend work due to pandemic prevention measures Keywords: remanufacturing system; component standardization; group technology; sorting algorithm; similarity coefficient 1. Introduction Industry 4.0 has profoundly changed manufacturing and remanufacturing. While the two systems differ, they both require finished goods to be delivered on time. Company leaders place tremendous importance on timely deliveries, but delays can be difficult to avoid. Two factors contribute to delays: a lack of materials and the low performance of manufacturing and remanufacturing systems. A shortage of materials can result when a company has too many types of components and not enough of each in stock. In other words, the more parts there are, the higher the risk of a shortage. Manufacturing and remanufacturing systems can perform poorly when workers are suddenly absent The two cases in this study consider a manufacturer whose major products are DC sources, AC sources, and power supplies. The R&D staff are divided into teams, with each responsible for different projects. This results in team A not knowing which components are used by team B, making component standardization more difficult. The vice president of manufacturing found that the products were not being delivered on time. One of the reasons was a lack of materials. More than 15,000 components were on the bill of materials, and not enough of each part was kept in stock. The vice president wanted to exchange one part for another, but if part i has not been approved by the R&D engineers, then it cannot be used to replace part j , even if the two are similar. The vice president wanted to reduce the number of components and increase the safety stock levels. In addition, the vice president wanted to rotate the workers so that the finished goods could be delivered on time, but the substituted workers were initially inefficient because some of them had never performed the tasks Sustainability 2021 , 13 , 7880. https://doi.org/10.3390/su 13147880 https://www.mdpi.com/journal/sustainability

Sustainability 2021 , 13 , 7880 2 of 17 Component standardization could solve the problem of the lack of materials, and the rotation of workers with similar skills could keep performance high even though workers were absent 2. Component Standardization in Manufacturing and Remanufacturing Systems 2.1. Manufacturing Traditionally, manufacturing is a process that transforms raw material into finished products to be sold to customers. The changing environment, especially an uncertain market, places company management under pressure to reduce costs and deliver products on time. The company that uses the manufacturing system best suited to its environment can increase its production efficiency. There are many types of manufacturing systems. Dorf and Kusiak [ 1 ] give four manufacturing systems: custom, intermittent, continuous, and flexible. Classifying manufacturing systems based on machining type gives dedicated machining systems, flexible machining systems, and reconfigurable machining systems [ 2 ]. The introduction of Industry 4.0 has brought new trends to the design of manufacturing systems. Three concepts—focused-flexibility manufacturing systems (FFMSs), reconfigurable manufacturing systems (RMSs), and smart manufacturing systems (SMSs)—have met the challenges of the Industry 4.0 philosophy [ 3 ]. The characteristics of manufacturing systems are uncertain due to the uncertainty inherent in customer demand, the supply chain, and the manufacturing process. There are at least two factors that increase the complexity of a manufacturing system: labor and the number of parts [ 4 ]. For example, a worker might be absent due to circumstances related to the COVID-19 pandemic, and the greater the number of parts, the greater the risk that some materials are lacking. One Industry 4.0 trend is advanced manufacturing systems [ 5 ]. Advanced manufacturing technology (AMT) has in part become critical due to the need for high-value, low-volume products [ 2 ]. As AMT requires workers with advanced skills, the effectiveness and efficiency of the manufacturing process and the quality of the goods produced depend on the skills of the workforce [ 6 ]. The training of workers has taken on new significance, with managers showing proficiency at sending select workers to attend essential classes on a limited budget 2.2. Remanufacturing In addition to waste disposal, three end-of-life strategies are available for industrial companies: reuse, remanufacturing, and recycling [ 7 ]. Supporting a sustainable environment on Earth is essential. Several product recovery options promote a sustainable environment. Recycling and remanufacturing are the two most advanced product recovery options [ 8 ]. Remanufacturing has become a prominent topic, especially in the last 20 years Because the concepts of product multilife cycle and extended producer responsibility have been popularized [ 9 , 10 ], remanufacturing has become an integral part of many supply chain operations [ 11 ]. Rather than being disposed of, numerous used products and their components can be reused. Reusing older product components or semiproducts and reinstating them back into the manufacturing process for new product production may be efficient policies for manufacturing enterprises. Remanufacturing is defined as “recycling by manufacturing ‘good as new’ products from used products” [ 7 ]. The key difference between manufacturing and remanufacturing is that manufacturing involves the use of new components, whereas remanufacturing involves the use of used components. In addition, the basic life cycle stages of the manufacturing system include design, sourcing materials, production, assembly, and quality control; the remanufacturing system stages include disassembly, sorting, inspection, cleaning, refurbishment, reassembly, and quality control [ 2 ]. Because the life cycle stages of the remanufacturing system include cleaning and refurbishment, parts that are renewable are used in the remanufacturing process. Hence, producing components from raw materials at the outset is not required, and, therefore, the

Sustainability 2021 , 13 , 7880 3 of 17 remanufacturing process reduces the energy consumed in the process of production and thus limits environmental damage [ 12 ]; it also prolongs the lifetime of products [ 13 ]. Apart from reduced energy consumption, the benefits of remanufacturing compared with traditional manufacturing are as follows: less consumption of raw materials [ 7 , 14 , 15 ], less pollution, and less required investment [ 2 ]. In addition, Das and Dutta’s [ 16 ] research indicated that product and component remanufacturing can reduce order variations and the bullwhip effect at both the retailer and distributor levels. In the life cycle of reusable products, especially in reverse logistics management or in a closed-loop supply chain, remanufacturing continues to play a critical role. Thus, the remanufacturing model is not only circular in nature but also economical [ 17 ]. Although remanufacturing can, compared with traditional manufacturing, reduce pollution, investment, and energy consumption, some risks arise. The major risk is uncertainty [ 2 , 18 , 19 ]. The uncertain quality and quantity associated with remanufacturing affect the confidence in and support for remanufacturing businesses. Industry 4.0 focuses on the integration of the Internet of Things (IoT). In the IoT, modern smart technology is used that offers solutions for remanufacturing. Fatimah et al. [ 20 ] used Industry 4.0 technologies to develop a sustainable waste management system that includes mixed collecting, sorting, transportation, varied treatment, and chained disposal functions. However, the problem of uncertain quality persists Sorting and inspection are vital processes of the remanufacturing system. Before the dispatch of reusable items to the remanufacturing chain, all waste items must be sorted and classified to determine which items can be recycled. Whether a component is to be reused or becomes waste depends on its quality. Yan et al. [ 21 ] demonstrated that a recycled product can be used as a second-hand product, repaired, or dissembled into materials according to its quality. Generally, for companies, the higher the ratio of reusable to disposable components, the greater is the economic profit 2.3. Component Standardization The cost of inventory in a remanufacturing system is higher than in a manufacturing system [ 2 ]. One resolution to this issue is component standardization. The term component standardization implies that several components can be replaced by a single component that has the specific functions of all the aforementioned components [ 22 ]. The traditional manufacturing system has less need for a costly inventory, and profits are obtained through bulk buying to achieve component standardization. In addition, component standardization can reduce labor costs and the requirement for multiskilled operators [ 22 ]. Achieving component standardization in the remanufacturing system can also reduce the number of material types, maximize homogeneous material recovery, and facilitate the disassembly and remanufacture of components [ 15 ]. Even the waste sorting step in the remanufacturing system can eventually become unnecessary [ 23 ]. However, component standardization has some disadvantages in the manufacturing and the remanufacturing system, such as resisting change, blocking progress, and favoring only large companies [ 24 ]. The standardized components might also have excess functionality [ 22 ] because the components must meet all of the requirements of the semiproduct or finished product that uses them The COVID-19 pandemic affected the global economy and supply chain considerably in 2020. However, retaliatory consumption triggered substantial growth in production demand when vaccines began to be deployed, and a resultant material and component shortage ensued. By implementing component standardization and stockpiling components earlier in preparation, such suboptimal situations could have been avoided. The benefits of component standardization are being leveraged in manufacturing systems based in one factory, but these benefits are being ignored in remanufacturing systems Thus, component standardization can reduce costs as well as the dependence on the parts provision in the reverse logistics process. In addition, component standardization can be

Sustainability 2021 , 13 , 7880 4 of 17 applied across different factories, enabling the recycling of parts at different locations if such parts have the same specification 2.4. Aim Companies need to handle the absence of workers and achieve component standardization. In the literature, the declarative model and constraint programming (CP) technology are used to resolve the problem of absent workers [ 25 , 26 ]. Szwarc et al. [ 27 ] even used this method to solve the teacher allocation problem. Several methods have been proposed to achieve component standardization, such as neural networks [ 28 ], the shape similarity algorithm [ 29 ], and the zero-based approach [ 30 ]. The aim of this study is to solve these two problems by using the concept of group technology (GT), both for manufacturing systems and remanufacturing systems 3. Methodology The algorithm discussed in this article is based on the concept of GT, especially in terms of the algorithm’s definition of the similarity coefficient 3.1. GT In a manufacturing or remanufacturing system, a manager must reduce relative costs and material handling. The cost of material handling constitutes 15% to 70% of the total manufacturing cost of a product [ 2 ]. The principal benefits associated with cellular manufacturing (CM) are reductions in throughput time, setup time, and inventory [ 31 , 32 ]. Hence, reductions in the overall processing time and material handling costs can be achieved through the optimum machine cell and part family formation [ 33 , 34 ]. The concept of GT has been implemented in CM [ 35 – 40 ]. GT algorithms can be used to group similar parts or machines [ 41 , 42 ], meaning that GT provides a method of sorting machines into machine cells and parts into part families [ 43 – 49 ]. Several algorithms have been developed to solve the component standardization problem, such as heuristics [ 50 – 55 ], genetic algorithms [ 56 – 60 ], and closed neighbor algorithms [ 43 , 44 , 61 ]. In addition, conventional GT was limited to two dimensions, but Parkin and Li expanded it to multidimensional GT in 1997 [ 44 ]. Conventional GT has the benefits of reducing material handling and part movement times. Consider the following incidence matrix A , where a 11 = 1 implies that p 1 must be operated on by m 1 , and a 21 = 0 implies that p 1 should not be operated on by m 2 A = p 1 p 2 p 3 p 4 m 1 m 2 m 3 m 4 m 5       1 1 1 1 1 1 1 1 1 1       If the plan of the incidence matrix is to let { m 1 , m 2 , m 3 , p 1 , p 2 } be a group and { m 4 , m 5 , p 3 , p 4 } be another group, then, for instance, part p 1 must be moved between two groups. However, when the GT algorithm is used in this example, two groups can be formed, and then all parts can be operated on in only one group without the requirement to move to another group Sustainability 2021 , 13 , x FOR PEER REVIEW 4 of 17 applied across different factories, enabling the recycling of parts at different locations if such parts have the same specification. 2.4. Aim Companies need to handle the absence of workers and achieve component standardization. In the literature, the declarative model and constraint programming (CP) technology are used to resolve the problem of absent workers [25,26]. Szwarc et al. [27] even used this method to solve the teacher allocation problem. Several methods have been proposed to achieve component standardization, such as neural networks [28], the shape similarity algorithm [29], and the zero-based approach [30]. The aim of this study is to solve these two problems by using the concept of group technology (GT), both for manufacturing systems and remanufacturing systems. 3. Methodology The algorithm discussed in this article is based on the concept of GT, especially in terms of the algorithm’s definition of the similarity coefficient 3.1. GT In a manufacturing or remanufacturing system, a manager must reduce relative costs and material handling. The cost of material handling constitutes 15% to 70% of the total manufacturing cost of a product [2]. The principal benefits associated with cellular manufacturing (CM) are reductions in throughput time, setup time, and inventory [31,32]. Hence, reductions in the overall processing time and material handling costs can be achieved through the optimum machine cell and part family formation [33,34]. The concept of GT has been implemented in CM [35–40]. GT algorithms can be used to group similar parts or machines [41,42], meaning that GT provides a method of sorting machines into machine cells and parts into part families [43–49]. Several algorithms have been developed to solve the component standardization problem, such as heuristics [50–55], genetic algorithms [56–60], and closed neighbor algorithms [43,44,61]. In addition, conventional GT was limited to two dimensions, but Parkin and Li expanded it to multidimensional GT in 1997 [44]. Conventional GT has the benefits of reducing material handling and part movement times. Consider the following incidence matrix {?} , where {?} = 1 implies that {?} must be operated on by {?} , and {?} = 0 implies that {?} should not be operated on by {?} . {?} {?} {?} {?} {?} = {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 1 1 1 1 ⎦ ⎥ ⎥ ⎥ ⎤ If the plan of the incidence matrix is to let { {?} , {?} , {?} , {?} , {?} } be a group and { {?} , {?} , {?} , {?} } be another group, then, for instance, part {?} must be moved between two groups. However, when the GT algorithm is used in this example, two groups can be formed, and then all parts can be operated on in only one group without the requirement to move to another group. {?} {?} {?} {?} {?} = {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 1 1 1 1 ⎦ ⎥ ⎥ ⎥ ⎤ Some authors have noted that the method of sorting different parts into a part family is based on the relationships between a part’s figure, operating process, and characteristics

Sustainability 2021 , 13 , 7880 5 of 17 Some authors have noted that the method of sorting different parts into a part family is based on the relationships between a part’s figure, operating process, and characteristics [ 62 ]. Won and Logendran [ 48 ] as well as Huang and Yan [ 63 ] have used a new similarity coefficient method to solve cell formation problems, and a typical similarity coefficient algorithm used for GT was proposed by McAuley [ 50 ]. sc ij = ε ij ε i + ε j − ε ij where ε i and ε j represent the number of parts processed by machines i and j , respectively, and ε ij represents the number of parts processed by both machines i and j The major obstacle in component standardization is the difficulty in ascertaining which parts can be standardized in large inventory data sets (bills of materials). Thus, initially, the groups include parts, and the products for which these parts are used must be identified The aim of component standardization can be identified through discussion with R&D engineers. A further problem in both manufacturing systems and remanufacturing systems is high labor costs [ 64 ]. This study can also solve this problem as explained in example 3 Conventional GT has two dimensions: machines and parts. However, the part figures, operating processes, and characteristics are the focus of this article, not machines To achieve component standardization, the following assumptions have been made: 1 The sequences of operations among parts are not considered 2 The volumes of parts are not considered Suppose an incidence matrix A exists with order m × n , which means it includes m machines and n parts. The incidence matrix is defined as follows: A = a ij , where a ij = 1 0 . If part j interacts with machine i , then a ij = 1; otherwise, a ij = 0 The parts used in a manufacturing process generally follow this sequence. Thus, the operational sequence of the incidence matrix is defined as follows: OS = os ij m × n where os ij is the os ij th operating process of part j The algorithm proposed in this study must transfer matrix OS to matrix A first. Thus, a ij = 1, if os ij 6 = 0 0, otherwise (1) 3.2. Machines Sorting The relationship between two machines in incidence matrix A is called the similarity coefficient and is defined as follows: SC m = sc ij m × m (2) where sc ij = ε ij ε i + ε j − ε ij (3) and ε ij = n ∑ k = 1 a ik a jk , i 6 = j ; ε kk = 0 ε i = n ∑ k = 1 a ik ε j = n ∑ k = 1 a jk Forming a machine cell involves six steps as follows: Step 1: Let the maximum similarity coefficient of each machine form the maximum similarity coefficient matrix, W = { w i } m × 1 , where w i = max { x | x = sc ik , 1 ≤ k ≤ m } .

Sustainability 2021 , 13 , 7880 6 of 17 Step 2: Form the ordering matrix O = { o i } m × 1 , where o i = m k and w k is the i th high value in matrix W . Ties are broken by choosing the larger number of w k values in row k of matrix SC m . If a tie occurs again, then the smaller k -value is chosen Step 3: Let the association set matrix, Λ = { λ i } m × 1 , λ i = { m x | sc m i m x = w i , 1 ≤ x ≤ m } Step 4: Let o 1 ∈ MC 1 Step 5: Let      o i ∈ MC j , i f ( o k ∈ λ o i ) V ∃ o k ∈ MC j A tie is broken by choosing the smallest n ( MC x ) o i ∈ MC y , otherwise , and MC y is a new machine cell (4) Step 6: Repeat step 5 until all machines are sorted 3.3. Parts Sorting Let SC n = sc ij n × n be the similarity coefficient between two parts in incidence matrix A . The parts sorting algorithm is similar to that for machines. The difference is that in this case, A T (transpose of incidence matrix A ) is used instead of A Now, MC m machine cells and PF n part families are present; thus, the final number of groups is as follows (although the value of MC m is generally equal to that of PF n ): n ( G i ) = min ( n ( MC m ) , n ( PF n )) I f MC m < PF n , PF x ⊂ MC y where ∑ a ij = max ∑ a αβ α ∈ MC σ , β ∈ PF x , i ∈ MC y , j ∈ PF x , 1 ≤ MC σ ≤ MC m A tie is broken by choosing the smallest group size otherwise , MC x ⊂ PF y where ∑ a ij = max ∑ a αβ α ∈ MC x , β ∈ PF σ , i ∈ MC x , j ∈ PF y , 1 ≤ PF σ ≤ PF n A tie is broken by choosing the smallest group size (5) 3.4. GT Efficiency One GT efficiency measurement equation was defined by Chandrasekharan and Rajagopalan [ 65 ] as follows. In this work, let q be equal to 0.5 η = q η 1 + ( 1 − q ) η 2 = q e b ∑ Q i P i + ( 1 − q ) 1 − e 0 mn − ∑ Q i P i (6) 4. Numerical Example and Case Studies Three examples were examined in this study: one traditional GT problem and two case studies. All three demonstrate how the algorithm proposed in this article can operate 4.1. A Comparison with Ahi’s Algorithm This example is taken from Ahi et al. [ 66 ], who referred to Boulif and Atif [ 67 ]. The original part–machine incidence matrix with consideration for the operational sequences of machines for producing parts is as follows: Sustainability 2021 , 13 , x FOR PEER REVIEW 7 of 17 {?} 1 {?} 2 {?} 3 {?} 4 {?} 5 {?} 6 {?} 7 {?} 8 {?} 9 {?} 1 0 {?} 1 1 {?} 1 2 {?} 1 3 {?} 1 4 {?} 1 5 {?} 1 6 {?} 1 7 {?} 1 8 {?} 1 9 {?} 2 0 {?}{?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 2 1 1 3 1 1 1 3 1 1 1 1 4 2 2 2 2 2 3 2 2 3 2 1 2 5 2 2 2 2 1 1 1 2 2 5 3 1 2 2 1 1 2 1 3 3 3 3 3 3 1 2 4 4 4 4 4 1 3 5 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ After Equation (1) is applied, the incidence matrix is obtained as follows: {?} 1 {?} 2 {?} 3 {?} 4 {?} 5 {?} 6 {?} 7 {?} 8 {?} 9 {?} 1 0 {?} 1 1 {?} 1 2 {?} 1 3 {?} 1 4 {?} 1 5 {?} 1 6 {?} 1 7 {?} 1 8 {?} 1 9 {?} 2 0 {?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ 4.1.1. Machine Sorting The relationship between the two machines must be obtained first. For instance, by using Equation (3), the relationship between machines {?} and {?} can be obtained as follows: {?}{?} = {?} {?} + {?} − {?} = 6 7 + 8 − 6 = 0.667 Thus, the similarity coefficient matrix {?}{?} is obtained as follows by using Equation (2). {?} {?} {?} {?} {?} {?} {?} {?} {?}{?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 0 0.133 0.9 0.059 0.063 0.067 0.125 0.067 0.133 0 0.067 0.667 0.077 0.083 0.667 0.857 0.9 0.067 0 0 0.067 0.071 0.063 0 0.059 0.667 0 0 0.25 0.167 0.6 0.75 0.063 0.077 0.067 0.25 0 0.444 0.154 0.083 0.067 0.083 0.071 0.167 0.444 0 0.167 0.091 0.125 0.667 0.063 0.6 0.154 0.167 0 0.75 0.067 0.857 0 0.75 0.083 0.091 0.75 0 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ Through step 1 to step 3 in Section 2.1, the maximum similarity coefficient matrix {?} , ordering matrix {?} , and association set matrix Λ are obtained as follows: {?} = ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 0.9 0.857 0.9 0.75 0.444 0.444 0.75 0.857⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ , {?} = ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ {?} {?} {?} {?} {?} {?} {?} {?} ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ , {?} = ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ {{?} } {{?} } {{?} } {{?} } {{?} } {{?} } {{?} } {{?} }⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ The first machine, belonging to the first machine cell ( {?}{?} ), is {?} in step 4. The next machine to be chosen is machine {?} , and it belongs to {?}{?} because {?} = {?} , and ({?} ∈ {?} ⋀ {?} ∈ {?}{?} ) in step 5 of using Equation (4). The third machine, belonging to the second

Sustainability 2021 , 13 , 7880 7 of 17 After Equation (1) is applied, the incidence matrix is obtained as follows: Sustainability 2021 , 13 , x FOR PEER REVIEW 7 of 17 {?} 1 {?} 2 {?} 3 {?} 4 {?} 5 {?} 6 {?} 7 {?} 8 {?} 9 {?} 1 0 {?} 1 1 {?} 1 2 {?} 1 3 {?} 1 4 {?} 1 5 {?} 1 6 {?} 1 7 {?} 1 8 {?} 1 9 {?} 2 0 {?}{?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 2 1 1 3 1 1 1 3 1 1 1 1 4 2 2 2 2 2 3 2 2 3 2 1 2 5 2 2 2 2 1 1 1 2 2 5 3 1 2 2 1 1 2 1 3 3 3 3 3 3 1 2 4 4 4 4 4 1 3 5 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ After Equation (1) is applied, the incidence matrix is obtained as follows: {?} 1 {?} 2 {?} 3 {?} 4 {?} 5 {?} 6 {?} 7 {?} 8 {?} 9 {?} 1 0 {?} 1 1 {?} 1 2 {?} 1 3 {?} 1 4 {?} 1 5 {?} 1 6 {?} 1 7 {?} 1 8 {?} 1 9 {?} 2 0 {?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ 4.1.1. Machine Sorting The relationship between the two machines must be obtained first. For instance, by using Equation (3), the relationship between machines {?} and {?} can be obtained as follows: {?}{?} = {?} {?} + {?} − {?} = 6 7 + 8 − 6 = 0.667 Thus, the similarity coefficient matrix {?}{?} is obtained as follows by using Equation (2). {?} {?} {?} {?} {?} {?} {?} {?} {?}{?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 0 0.133 0.9 0.059 0.063 0.067 0.125 0.067 0.133 0 0.067 0.667 0.077 0.083 0.667 0.857 0.9 0.067 0 0 0.067 0.071 0.063 0 0.059 0.667 0 0 0.25 0.167 0.6 0.75 0.063 0.077 0.067 0.25 0 0.444 0.154 0.083 0.067 0.083 0.071 0.167 0.444 0 0.167 0.091 0.125 0.667 0.063 0.6 0.154 0.167 0 0.75 0.067 0.857 0 0.75 0.083 0.091 0.75 0 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ Through step 1 to step 3 in Section 2.1, the maximum similarity coefficient matrix {?} , ordering matrix {?} , and association set matrix Λ are obtained as follows: {?} = ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 0.9 0.857 0.9 0.75 0.444 0.444 0.75 0.857⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ , {?} = ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ {?} {?} {?} {?} {?} {?} {?} {?} ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ , {?} = ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ {{?} } {{?} } {{?} } {{?} } {{?} } {{?} } {{?} } {{?} }⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ The first machine, belonging to the first machine cell ( {?}{?} ), is {?} in step 4. The next machine to be chosen is machine {?} , and it belongs to {?}{?} because {?} = {?} , and ({?} ∈ {?} ⋀ {?} ∈ {?}{?} ) in step 5 of using Equation (4). The third machine, belonging to the second 4.1.1. Machine Sorting The relationship between the two machines must be obtained first. For instance, by using Equation (3), the relationship between machines m 2 and m 4 can be obtained as follows: sc 24 = ε 24 ε 2 + ε 4 − ε 24 = 6 7 + 8 − 6 = 0.667 Thus, the similarity coefficient matrix SC m is obtained as follows by using Equation (2) Sustainability 2021 , 13 , x FOR PEER REVIEW 7 of 17 {?} 1 {?} 2 {?} 3 {?} 4 {?} 5 {?} 6 {?} 7 {?} 8 {?} 9 {?} 1 0 {?} 1 1 {?} 1 2 {?} 1 3 {?} 1 4 {?} 1 5 {?} 1 6 {?} 1 7 {?} 1 8 {?} 1 9 {?} 2 0 {?}{?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 2 1 1 3 1 1 1 3 1 1 1 1 4 2 2 2 2 2 3 2 2 3 2 1 2 5 2 2 2 2 1 1 1 2 2 5 3 1 2 2 1 1 2 1 3 3 3 3 3 3 1 2 4 4 4 4 4 1 3 5 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ After Equation (1) is applied, the incidence matrix is obtained as follows: {?} 1 {?} 2 {?} 3 {?} 4 {?} 5 {?} 6 {?} 7 {?} 8 {?} 9 {?} 1 0 {?} 1 1 {?} 1 2 {?} 1 3 {?} 1 4 {?} 1 5 {?} 1 6 {?} 1 7 {?} 1 8 {?} 1 9 {?} 2 0 {?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ 4.1.1. Machine Sorting The relationship between the two machines must be obtained first. For instance, by using Equation (3), the relationship between machines {?} and {?} can be obtained as follows: {?}{?} = {?} {?} + {?} − {?} = 6 7 + 8 − 6 = 0.667 Thus, the similarity coefficient matrix {?}{?} is obtained as follows by using Equation (2). {?} {?} {?} {?} {?} {?} {?} {?} {?}{?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 0 0.133 0.9 0.059 0.063 0.067 0.125 0.067 0.133 0 0.067 0.667 0.077 0.083 0.667 0.857 0.9 0.067 0 0 0.067 0.071 0.063 0 0.059 0.667 0 0 0.25 0.167 0.6 0.75 0.063 0.077 0.067 0.25 0 0.444 0.154 0.083 0.067 0.083 0.071 0.167 0.444 0 0.167 0.091 0.125 0.667 0.063 0.6 0.154 0.167 0 0.75 0.067 0.857 0 0.75 0.083 0.091 0.75 0 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ Through step 1 to step 3 in Section 2.1, the maximum similarity coefficient matrix {?} , ordering matrix {?} , and association set matrix Λ are obtained as follows: {?} = ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 0.9 0.857 0.9 0.75 0.444 0.444 0.75 0.857⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ , {?} = ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ {?} {?} {?} {?} {?} {?} {?} {?} ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ , {?} = ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ {{?} } {{?} } {{?} } {{?} } {{?} } {{?} } {{?} } {{?} }⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ The first machine, belonging to the first machine cell ( {?}{?} ), is {?} in step 4. The next machine to be chosen is machine {?} , and it belongs to {?}{?} because {?} = {?} , and ({?} ∈ {?} ⋀ {?} ∈ {?}{?} ) in step 5 of using Equation (4). The third machine, belonging to the second Through step 1 to step 3 in Section 2.1 , the maximum similarity coefficient matrix W , ordering matrix O , and association set matrix Λ are obtained as follows: W =             0.9 0.857 0.9 0.75 0.444 0.444 0.75 0.857             , O =             m 1 m 3 m 2 m 8 m 4 m 7 m 5 m 6             , Λ =             { m 3 } { m 8 } { m 1 } { m 8 } { m 6 } { m 5 } { m 8 } { m 2 }             The first machine, belonging to the first machine cell ( MC 1 ), is m 1 in step 4. The next machine to be chosen is machine m 3 , and it belongs to MC 1 because o 2 = m 3 , and ( m 1 ∈ λ 3 V m 1 ∈ MC 1 ) in step 5 of using Equation (4). The third machine, belonging to the second machine cell ( MC 2 ), is m 2 . This is because o 3 = m 2 , and ( m 8 ∈ λ 2 V m 8 / ∈ MC 1 ) ; thus, m 2 must belong to a new machine cell called MC 2 in step 5 of Equation (4) Finally, three machine cells are present: MC 1 = { m 1 , m 3 } , MC 2 = { m 2 , m 8 , m 4 , m 7 } , and MC 3 = { m 5 , m 6 } .

Sustainability 2021 , 13 , 7880 8 of 17 4.1.2. Part Sorting First, the transpose of matrix A ( A T ) instead of A is ascertained as follows: Sustainability 2021 , 13 , x FOR PEER REVIEW 8 of 17 machine cell ( {?}{?} ), is {?} . This is because {?} = {?} , and ({?} ∈ {?} ⋀ {?} ∉ {?}{?} ) ; thus, {?} must belong to a new machine cell called {?}{?} in step 5 of Equation (4). Finally, three machine cells are present: {?}{?} = {{?} , {?} } , {?}{?} = {{?} , {?} , {?} , {?} } , and {?}{?} = {{?} , {?} } . 4.1.2. Part Sorting First, the transpose of matrix {?} ( {?} ) instead of {?} is ascertained as follows: {?} {?} {?} {?} {?} {?} {?} {?} {?} = {?} 1 {?} 2 {?} 3 {?} 4 {?} 5 {?} 6 {?} 7 {?} 8 {?} 9 {?} 10 {?} 11 {?} 12 {?} 13 {?} 14 {?} 15 {?} 16 {?} 17 {?} 18 {?} 19 {?} 20 ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ Through step 1 to step 3 in Section 2.1, the maximum similarity coefficient matrix {?} , ordering matrix {?} , and association set matrix Λ are obtained as follows: {?} = ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 0.8 1 1 0.8 1 1 0.667 0.667 0.667 0.667 1 0.667 0.667 1 0.667 1 1 0.8 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ , {?} = ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ , {?} = ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ {{?} } {{?} , {?} , {?} , {?} } {{?} , {?} , {?} } {{?} , {?} } {{?} , {?} } {{?} , {?} } {{?} } { {?} , {?} , {?} } {{?} , {?} , {?} , {?} , {?} } {{?} , {?} , {?} } {{?} , {?} , {?} , {?} , {?} } {{?} , {?} } {{?} , {?} , {?} , {?} } {{?} , {?} , {?} , {?} , {?} } {{?} } {{?} , {?} , {?} , {?} } {{?} , {?} , {?} , {?} , {?} } {{?} , {?} } {{?} , {?} , {?} , {?} } {{?} , {?} , {?} } ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ Thus, by using the same method, three part families are identified: {?}{?} = {{?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} } , {?}{?} = {{?} , {?} , {?} , {?} , {?} , {?} } , and {?}{?} = {{?} , {?} , {?} , {?} , {?} } . Finally, the three groups are ordered as detailed in Table 1. Through step 1 to step 3 in Section 2.1 , the maximum similarity coefficient matrix W , ordering matrix O , and association set matrix Λ are obtained as follows: W =                                       1 1 0.8 1 1 0.8 1 1 0.667 0.667 0.667 0.667 1 0.667 0.667 1 0.667 1 1 0.8                                       , O =                                       p 2 p 8 p 13 p 16 p 19 p 4 p 7 p 18 p 1 p 5 p 3 p 6 p 20 p 9 p 11 p 14 p 17 p 10 p 12 p 15                                       , Λ =                                       { p 5 } { p 8 , p 13 , p 16 , p 19 } { p 4 , p 7 , p 18 } { p 7 , p 18 } { p 10 , p 12 } { p 7 , p 18 } { p 18 } { p 13 , p 16 , p 19 } { p 2 , p 8 , p 13 , p 16 , p 19 } { p 1 , p 5 , p 15 } { p 2 , p 8 , p 13 , p 16 , p 19 } { p 1 , p 5 } { p 2 , p 8 , p 16 , p 19 } { p 2 , p 8 , p 13 , p 16 , p 19 } { p 10 } { p 2 , p 8 , p 13 , p 19 } { p 2 , p 8 , p 13 , p 16 , p 19 } { p 4 , p 7 } { p 2 , p 8 , p 13 , p 16 } { p 4 , p 7 , p 18 }                                       Thus, by using the same method, three part families are identified: PF 1 = { p 2 , p 8 , p 13 , p 16 , p 19 , p 9 , p 11 , p 14 , p 17 } , PF 2 = { p 4 , p 7 , p 18 , p 3 , p 6 , p 20 } , and PF 3 = { p 1 , p 5 , p 10 , p 12 , p 15 } . Finally, the three groups are ordered as detailed in Table 1 .

Sustainability 2021 , 13 , 7880 9 of 17 Table 1. Three groups in example 1 Group Member in the Group G 1 m 1 , m 3 , p 2 , p 8 , p 13 , p 16 , p 19 , p 9 , p 11 , p 14 , p 17 G 2 m 2 , m 8 , m 4 , m 7 , p 4 , p 7 , p 18 , p 3 , p 6 , p 20 , G 3 m 5 , m 6 , p 1 , p 5 , p 10 , p 12 , p 15 The sorted incidence matrix and sorted operational sequences of the incidence matrix are as follows Sustainability 2021 , 13 , x FOR PEER REVIEW 9 of 17 Table 1. Three groups in example 1. Group Member in the Group {?} {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} {?} {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} {?} {?} , {?} , {?} , {?} , {?} , {?} , {?} The sorted incidence matrix and sorted operational sequences of the incidence matrix are as follows. {?} 2 {?} 8 {?} 1 3 {?} 1 6 {?} 1 9 {?} 9 {?} 1 1 {?} 1 4 {?} 1 7 {?} 4 {?} 7 {?} 1 8 {?} 3 {?} 6 {?} 2 0 {?} 1 {?} 5 {?} 1 0 {?} 1 2 {?} 1 5 {?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ {?} 2 {?} 8 {?} 1 3 {?} 1 6 {?} 1 9 {?} 9 {?} 1 1 {?} 1 4 {?} 1 7 {?} 4 {?} 7 {?} 1 8 {?} 3 {?} 6 {?} 2 0 {?} 1 {?} 5 {?} 1 0 {?} 1 2 {?} 1 5 {?}{?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 3 1 3 2 2 2 2 2 2 3 2 3 1 2 1 4 2 1 1 2 4 1 3 4 4 5 2 2 1 5 2 1 2 1 1 3 3 4 3 3 4 2 2 5 2 2 3 1 2 2 3 1 1 1 3 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ The final result obtained using the algorithm introduced by Ahi et al. [66] is as follows: {?} 2 {?} 8 {?} 1 3 {?} 1 6 {?} 1 9 {?} 1 7 {?} 9 {?} 1 {?} 5 {?} 1 5 {?} 1 1 {?} 1 0 {?} 1 4 {?} 1 2 {?} 4 {?} 1 8 {?} 7 {?} 3 {?} 6 {?} 2 0 {?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ {?} 2 {?} 8 {?} 1 3 {?} 1 6 {?} 1 9 {?} 1 7 {?} 9 {?} 1 {?} 5 {?} 1 5 {?} 1 1 {?} 1 0 {?} 1 4 {?} 1 2 {?} 4 {?} 1 8 {?} 7 {?} 3 {?} 6 {?} 2 0 {?}{?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 4 3 1 4 4 5 2 1 2 4 1 1 2 1 2 3 4 3 3 3 4 1 2 2 1 2 5 2 1 2 1 1 1 3 3 2 2 2 2 3 1 5 2 2 2 2 2 1 3 2 3 1 1 1 1 1 3 1 3 1 2 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ However, two parts— {?} and {?} —belong to {?} ; {?} and {?} do not operate on these two parts. The GT efficiency of the algorithm proposed in this study is obtained using Equation (6), as follows: In the sorted incidence matrix {?} with order 8 × 20, {?} = 18 + 24 + 9 because there are eighteen 1 s in group 1, twenty-four 1 s in group 2, and nine 1 s in group 3. ∑ {?} {?} = 2 ∙ 9 + 4 ∙ 6 + 2 ∙ 5 because the order of these three groups are 9 × 2, 4 × 6, and 2 × 5, respectively. {?} = 10 because there are ten 1 s not in these three groups. Sustainability 2021 , 13 , x FOR PEER REVIEW 9 of 17 Table 1. Three groups in example 1. Group Member in the Group {?} {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} {?} {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} {?} {?} , {?} , {?} , {?} , {?} , {?} , {?} The sorted incidence matrix and sorted operational sequences of the incidence matrix are as follows. {?} 2 {?} 8 {?} 1 3 {?} 1 6 {?} 1 9 {?} 9 {?} 1 1 {?} 1 4 {?} 1 7 {?} 4 {?} 7 {?} 1 8 {?} 3 {?} 6 {?} 2 0 {?} 1 {?} 5 {?} 1 0 {?} 1 2 {?} 1 5 {?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ {?} 2 {?} 8 {?} 1 3 {?} 1 6 {?} 1 9 {?} 9 {?} 1 1 {?} 1 4 {?} 1 7 {?} 4 {?} 7 {?} 1 8 {?} 3 {?} 6 {?} 2 0 {?} 1 {?} 5 {?} 1 0 {?} 1 2 {?} 1 5 {?}{?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 3 1 3 2 2 2 2 2 2 3 2 3 1 2 1 4 2 1 1 2 4 1 3 4 4 5 2 2 1 5 2 1 2 1 1 3 3 4 3 3 4 2 2 5 2 2 3 1 2 2 3 1 1 1 3 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ The final result obtained using the algorithm introduced by Ahi et al. [66] is as follows: {?} 2 {?} 8 {?} 1 3 {?} 1 6 {?} 1 9 {?} 1 7 {?} 9 {?} 1 {?} 5 {?} 1 5 {?} 1 1 {?} 1 0 {?} 1 4 {?} 1 2 {?} 4 {?} 1 8 {?} 7 {?} 3 {?} 6 {?} 2 0 {?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ {?} 2 {?} 8 {?} 1 3 {?} 1 6 {?} 1 9 {?} 1 7 {?} 9 {?} 1 {?} 5 {?} 1 5 {?} 1 1 {?} 1 0 {?} 1 4 {?} 1 2 {?} 4 {?} 1 8 {?} 7 {?} 3 {?} 6 {?} 2 0 {?}{?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 4 3 1 4 4 5 2 1 2 4 1 1 2 1 2 3 4 3 3 3 4 1 2 2 1 2 5 2 1 2 1 1 1 3 3 2 2 2 2 3 1 5 2 2 2 2 2 1 3 2 3 1 1 1 1 1 3 1 3 1 2 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ However, two parts— {?} and {?} —belong to {?} ; {?} and {?} do not operate on these two parts. The GT efficiency of the algorithm proposed in this study is obtained using Equation (6), as follows: In the sorted incidence matrix {?} with order 8 × 20, {?} = 18 + 24 + 9 because there are eighteen 1 s in group 1, twenty-four 1 s in group 2, and nine 1 s in group 3. ∑ {?} {?} = 2 ∙ 9 + 4 ∙ 6 + 2 ∙ 5 because the order of these three groups are 9 × 2, 4 × 6, and 2 × 5, respectively. {?} = 10 because there are ten 1 s not in these three groups. The final result obtained using the algorithm introduced by Ahi et al. [ 66 ] is as follows: Sustainability 2021 , 13 , x FOR PEER REVIEW 9 of 17 Table 1. Three groups in example 1. Group Member in the Group {?} {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} {?} {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} {?} {?} , {?} , {?} , {?} , {?} , {?} , {?} The sorted incidence matrix and sorted operational sequences of the incidence matrix are as follows. {?} 2 {?} 8 {?} 1 3 {?} 1 6 {?} 1 9 {?} 9 {?} 1 1 {?} 1 4 {?} 1 7 {?} 4 {?} 7 {?} 1 8 {?} 3 {?} 6 {?} 2 0 {?} 1 {?} 5 {?} 1 0 {?} 1 2 {?} 1 5 {?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ {?} 2 {?} 8 {?} 1 3 {?} 1 6 {?} 1 9 {?} 9 {?} 1 1 {?} 1 4 {?} 1 7 {?} 4 {?} 7 {?} 1 8 {?} 3 {?} 6 {?} 2 0 {?} 1 {?} 5 {?} 1 0 {?} 1 2 {?} 1 5 {?}{?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 3 1 3 2 2 2 2 2 2 3 2 3 1 2 1 4 2 1 1 2 4 1 3 4 4 5 2 2 1 5 2 1 2 1 1 3 3 4 3 3 4 2 2 5 2 2 3 1 2 2 3 1 1 1 3 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ The final result obtained using the algorithm introduced by Ahi et al. [66] is as follows: {?} 2 {?} 8 {?} 1 3 {?} 1 6 {?} 1 9 {?} 1 7 {?} 9 {?} 1 {?} 5 {?} 1 5 {?} 1 1 {?} 1 0 {?} 1 4 {?} 1 2 {?} 4 {?} 1 8 {?} 7 {?} 3 {?} 6 {?} 2 0 {?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ {?} 2 {?} 8 {?} 1 3 {?} 1 6 {?} 1 9 {?} 1 7 {?} 9 {?} 1 {?} 5 {?} 1 5 {?} 1 1 {?} 1 0 {?} 1 4 {?} 1 2 {?} 4 {?} 1 8 {?} 7 {?} 3 {?} 6 {?} 2 0 {?}{?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 4 3 1 4 4 5 2 1 2 4 1 1 2 1 2 3 4 3 3 3 4 1 2 2 1 2 5 2 1 2 1 1 1 3 3 2 2 2 2 3 1 5 2 2 2 2 2 1 3 2 3 1 1 1 1 1 3 1 3 1 2 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ However, two parts— {?} and {?} —belong to {?} ; {?} and {?} do not operate on these two parts. The GT efficiency of the algorithm proposed in this study is obtained using Equation (6), as follows: In the sorted incidence matrix {?} with order 8 × 20, {?} = 18 + 24 + 9 because there are eighteen 1 s in group 1, twenty-four 1 s in group 2, and nine 1 s in group 3. ∑ {?} {?} = 2 ∙ 9 + 4 ∙ 6 + 2 ∙ 5 because the order of these three groups are 9 × 2, 4 × 6, and 2 × 5, respectively. {?} = 10 because there are ten 1 s not in these three groups. Sustainability 2021 , 13 , x FOR PEER REVIEW 9 of 17 Table 1. Three groups in example 1. Group Member in the Group {?} {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} {?} {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} {?} {?} , {?} , {?} , {?} , {?} , {?} , {?} The sorted incidence matrix and sorted operational sequences of the incidence matrix are as follows. {?} 2 {?} 8 {?} 1 3 {?} 1 6 {?} 1 9 {?} 9 {?} 1 1 {?} 1 4 {?} 1 7 {?} 4 {?} 7 {?} 1 8 {?} 3 {?} 6 {?} 2 0 {?} 1 {?} 5 {?} 1 0 {?} 1 2 {?} 1 5 {?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ {?} 2 {?} 8 {?} 1 3 {?} 1 6 {?} 1 9 {?} 9 {?} 1 1 {?} 1 4 {?} 1 7 {?} 4 {?} 7 {?} 1 8 {?} 3 {?} 6 {?} 2 0 {?} 1 {?} 5 {?} 1 0 {?} 1 2 {?} 1 5 {?}{?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 3 1 3 2 2 2 2 2 2 3 2 3 1 2 1 4 2 1 1 2 4 1 3 4 4 5 2 2 1 5 2 1 2 1 1 3 3 4 3 3 4 2 2 5 2 2 3 1 2 2 3 1 1 1 3 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ The final result obtained using the algorithm introduced by Ahi et al. [66] is as follows: {?} 2 {?} 8 {?} 1 3 {?} 1 6 {?} 1 9 {?} 1 7 {?} 9 {?} 1 {?} 5 {?} 1 5 {?} 1 1 {?} 1 0 {?} 1 4 {?} 1 2 {?} 4 {?} 1 8 {?} 7 {?} 3 {?} 6 {?} 2 0 {?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ {?} 2 {?} 8 {?} 1 3 {?} 1 6 {?} 1 9 {?} 1 7 {?} 9 {?} 1 {?} 5 {?} 1 5 {?} 1 1 {?} 1 0 {?} 1 4 {?} 1 2 {?} 4 {?} 1 8 {?} 7 {?} 3 {?} 6 {?} 2 0 {?}{?} = {?} {?} {?} {?} {?} {?} {?} {?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 4 3 1 4 4 5 2 1 2 4 1 1 2 1 2 3 4 3 3 3 4 1 2 2 1 2 5 2 1 2 1 1 1 3 3 2 2 2 2 3 1 5 2 2 2 2 2 1 3 2 3 1 1 1 1 1 3 1 3 1 2 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ However, two parts— {?} and {?} —belong to {?} ; {?} and {?} do not operate on these two parts. The GT efficiency of the algorithm proposed in this study is obtained using Equation (6), as follows: In the sorted incidence matrix {?} with order 8 × 20, {?} = 18 + 24 + 9 because there are eighteen 1 s in group 1, twenty-four 1 s in group 2, and nine 1 s in group 3. ∑ {?} {?} = 2 ∙ 9 + 4 ∙ 6 + 2 ∙ 5 because the order of these three groups are 9 × 2, 4 × 6, and 2 × 5, respectively. {?} = 10 because there are ten 1 s not in these three groups.

Sustainability 2021 , 13 , 7880 10 of 17 However, two parts— p 11 and p 14 —belong to G 2 ; m 5 and m 6 do not operate on these two parts The GT efficiency of the algorithm proposed in this study is obtained using Equation (6), as follows: In the sorted incidence matrix A sort with order 8 × 20, e b = 18 + 24 + 9 because there are eighteen 1 s in group 1, twenty-four 1 s in group 2, and nine 1 s in group 3 ∑ Q i P i = 2 · 9 + 4 · 6 + 2 · 5 because the order of these three groups are 9 × 2, 4 × 6, and 2 × 5, respectively e 0 = 10 because there are ten 1 s not in these three groups Thus, η = 0.5 · 18 + 24 + 9 2 · 9 + 4 · 6 + 2 · 5 + 0.5 · 1 − 10 8 · 20 − ( 2 · 9 + 4 · 6 + 2 · 5 ) = 0.944 In the same way, GT efficiency is calculated using Ahi’s algorithm, as follows: η = 0.5 · 14 + 9 + 24 2 · 7 + 2 · 7 + 4 · 6 + 0.5 · 1 − 14 8 · 20 − ( 2 · 7 + 2 · 7 + 4 · 6 ) = 0.887 4.2. Case Study of Component Standardization To solve the component standardization problem in a manufacturing or remanufacturing system by using the GT algorithm, the first step is to identify the product, the figure of parts, operating process of parts, or characteristics of parts (in substitute for machines) in a part–machine incidence matrix. This example illustrates how to locate the relevant components and then achieve component standardization. Eleven products or semiproducts and 12 parts are chosen. The parts are wires, and the specification is American wire gauge, as indicated in Table 2 . Table 2. Specification of parts Part Spec. Length (mm) p 1 AWG 16 7.5 p 2 AWG 21 10.0 p 3 AWG 13 12.0 p 4 AWG 13 5.5 p 5 AWG 20 13.5 p 6 AWG 17 8.0 p 7 AWG 19 10.5 p 8 AWG 14 12.0 p 9 AWG 16 10.0 p 10 AWG 18 8.0 p 11 AWG 15 13.0 p 12 AWG 17 9.5

Sustainability 2021 , 13 , 7880 11 of 17 The incidence matrix is as follows: Sustainability 2021 , 13 , x FOR PEER REVIEW 10 of 17 Thus, {?} = 0.5 ∙ 18 + 24 + 9 2 ∙ 9 + 4 ∙ 6 + 2 ∙ 5 + 0.5 ∙ 1 − 10 8 ∙ 20 − 2 ∙ 9 + 4 ∙ 6 + 2 ∙ 5 = 0.944 In the same way, GT efficiency is calculated using Ahi’s algorithm, as follows: {?} = 0.5 ∙ 14 + 9 + 24 2 ∙ 7 + 2 ∙ 7 + 4 ∙ 6 + 0.5 ∙ 1 − 14 8 ∙ 20 − 2 ∙ 7 + 2 ∙ 7 + 4 ∙ 6 = 0.887 4.2. Case Study of Component Standardization To solve the component standardization problem in a manufacturing or remanufacturing system by using the GT algorithm, the first step is to identify the product, the figure of parts, operating process of parts, or characteristics of parts (in substitute for machines) in a part–machine incidence matrix. This example illustrates how to locate the relevant components and then achieve component standardization. Eleven products or semiproducts and 12 parts are chosen. The parts are wires, and the specification is American wire gauge, as indicated in Table 2. Table 2. Specification of parts. Part Spec. Length (mm) {?} AWG 16 7.5 {?} AWG 21 10.0 {?} AWG 13 12.0 {?} AWG 13 5.5 {?} AWG 20 13.5 {?} AWG 17 8.0 {?} AWG 19 10.5 {?} AWG 14 12.0 {?} AWG 16 10.0 {?} AWG 18 8.0 {?} AWG 15 13.0 {?} AWG 17 9.5 The incidence matrix is as follows: {?} 1 {?} 2 {?} 3 {?} 4 {?} 5 {?} 6 {?} 7 {?} 8 {?} 9 {?} 1 0 {?} 1 1 {?} 1 2 {?} = {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ The groups are sorted by using the algorithm proposed in this article, as indicated in Table 3 , and four groups were categorized in the sorted incidence matrix as follows: Sustainability 2021 , 13 , x FOR PEER REVIEW 11 of 17 The groups are sorted by using the algorithm proposed in this article, as indicated in Table 3, and four groups were categorized in the sorted incidence matrix as follows: {?} 2 {?} 7 {?} 1 0 {?} 1 {?} 5 {?} 1 2 {?} 6 {?} 9 {?} 3 {?} 1 1 {?} 4 {?} 8 {?} = {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ Table 3. Four groups in example 2. Group Member in the Group {?} {?} , {?} , {?} , {?}{?}{?} , {?}{?}{?} {?} {?} , {?} , {?} , {?} , {?}{?}{?} , {?}{?}{?} , {?}{?}{?} {?} {?} , {?} , {?} , {?}{?}{?} , {?}{?}{?} , {?}{?}{?} {?} {?} , {?} , {?}{?}{?} , {?}{?}{?} , {?}{?}{?} 4.3. Case Study of a Printed Circuit Board Assembly Line As mentioned in Section 3.1, manufacturing and remanufacturing systems are subject to the problem of high labor costs and absent workers. The problem of worker rotation can also be solved using the algorithm proposed in this study. Consider that the factory outlined in the previous example contains a printed circuit board (PCB) assembly line that includes surface mount technology and insertion processes. Workers were absent due to COVID-19. The major operations of a PCB assembly line are divided among 14 workstations, as described in Table 4. Table 4. Fourteen workstations. {?}{?} Solder paste printing {?}{?} Glue dispensing {?}{?} High-speed placement {?}{?} Multifunctional mounter {?}{?} Manual insertion for surface mount technology {?}{?} Reflow {?}{?} Visual inspection {?}{?} Auto-optical inspection {?}{?} Axial insertion {?}{?} Radial insertion {?}{?} Dual in-line package insertion {?}{?} Manual insertion for general insertion {?}{?} Wave soldering {?}{?} In-circuit test In addition, 23 operators are working on this assembly line. The incidence matrix is as follows, but the workstations are substituted for machines and operators substituted for parts. In this incidence matrix, entry “1” implies that the worker can operate the process or has previously undertaken the relevant training course involved with the workstation. Table 3. Four groups in example 2 Group Member in the Group G 1 p 2 , p 7 , p 10 , pdt 6 , pdt 2 G 2 p 1 , p 5 , p 12 , p 6 , pdt 8 , pdt 5 , pdt 10 G 3 p 9 , p 3 , p 11 , pdt 1 , pdt 7 , pdt 11 G 4 p 4 , p 8 , pdt 3 , pdt 9 , pdt 4 4.3. Case Study of a Printed Circuit Board Assembly Line As mentioned in Section 3.1 , manufacturing and remanufacturing systems are subject to the problem of high labor costs and absent workers. The problem of worker rotation can also be solved using the algorithm proposed in this study Consider that the factory outlined in the previous example contains a printed circuit board (PCB) assembly line that includes surface mount technology and insertion processes Workers were absent due to COVID-19. The major operations of a PCB assembly line are divided among 14 workstations, as described in Table 4 .

Sustainability 2021 , 13 , 7880 12 of 17 Table 4. Fourteen workstations wk 1 Solder paste printing wk 2 Glue dispensing wk 3 High-speed placement wk 4 Multifunctional mounter wk 5 Manual insertion for surface mount technology wk 6 Reflow wk 7 Visual inspection wk 8 Auto-optical inspection wk 9 Axial insertion wk 10 Radial insertion wk 11 Dual in-line package insertion wk 12 Manual insertion for general insertion wk 13 Wave soldering wk 14 In-circuit test In addition, 23 operators are working on this assembly line. The incidence matrix is as follows, but the workstations are substituted for machines and operators substituted for parts. In this incidence matrix, entry “1” implies that the worker can operate the process or has previously undertaken the relevant training course involved with the workstation Sustainability 2021 , 13 , x FOR PEER REVIEW 12 of 17 {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} = {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ The five groups generated using the proposed algorithm are specified in Table 5 as follows: Table 5. Five groups in example 3. Group Member in the Group {?} {?}{?} , {?}{?} , {?}{?} , {?}{?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} {?} {?}{?} , {?}{?} , {?} , {?} , {?} , {?} , {?} , {?} {?} {?}{?} , {?}{?} , {?}{?} , {?} , {?} {?} {?}{?} , {?}{?} , {?} , {?} , {?} , {?} {?} {?}{?} , {?}{?} , {?}{?} , {?} , {?} , {?} The sorted incidence matrix is as follows. {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} = {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ 5. Results and Discussion In the numerical example, the algorithm proposed in this study is used to categorize three groups, as detailed in Section 4.1. Ahi‘s algorithm was compared to the GT efficiency measure proposed by Chandrasekharan and Rajagopalan. The GT efficiency of the algorithm proposed in this study is 0.944, higher than the 0.887 shown in the literature for Ahi’s algorithm. Thus, the GT algorithm proposed in this study works and can perform better than some algorithms in the literature. All parts in each group are to be operated on by at least one machine that is in the same group; this specification is applied using the proposed algorithm in this study. In Ahi’s algorithm, however, the two parts {?} and {?} belong to group {?} , but all machines ( {?} and {?} ) that are in the same group do not operate on these two parts. The five groups generated using the proposed algorithm are specified in Table 5 as follows: Table 5. Five groups in example 3 Group Member in the Group G 1 wk 1 , wk 2 , wk 6 , wk 5 , o 10 , o 17 , o 6 , o 19 , o 1 , o 3 , o 20 , o 21 G 2 wk 12 , wk 13 , o 5 , o 22 , o 4 , o 9 , o 14 , o 16 G 3 wk 7 , wk 14 , wk 11 , o 8 , o 11 G 4 wk 9 , wk 10 , o 13 , o 15 , o 2 , o 18 G 5 wk 3 , wk 4 , wk 8 , o 7 , o 12 , o 23

Sustainability 2021 , 13 , 7880 13 of 17 The sorted incidence matrix is as follows Sustainability 2021 , 13 , x FOR PEER REVIEW 12 of 17 {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} = {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ The five groups generated using the proposed algorithm are specified in Table 5 as follows: Table 5. Five groups in example 3. Group Member in the Group {?} {?}{?} , {?}{?} , {?}{?} , {?}{?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} , {?} {?} {?}{?} , {?}{?} , {?} , {?} , {?} , {?} , {?} , {?} {?} {?}{?} , {?}{?} , {?}{?} , {?} , {?} {?} {?}{?} , {?}{?} , {?} , {?} , {?} , {?} {?} {?}{?} , {?}{?} , {?}{?} , {?} , {?} , {?} The sorted incidence matrix is as follows. {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} = {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ 5. Results and Discussion In the numerical example, the algorithm proposed in this study is used to categorize three groups, as detailed in Section 4.1. Ahi‘s algorithm was compared to the GT efficiency measure proposed by Chandrasekharan and Rajagopalan. The GT efficiency of the algorithm proposed in this study is 0.944, higher than the 0.887 shown in the literature for Ahi’s algorithm. Thus, the GT algorithm proposed in this study works and can perform better than some algorithms in the literature. All parts in each group are to be operated on by at least one machine that is in the same group; this specification is applied using the proposed algorithm in this study. In Ahi’s algorithm, however, the two parts {?} and {?} belong to group {?} , but all machines ( {?} and {?} ) that are in the same group do not operate on these two parts. 5. Results and Discussion In the numerical example, the algorithm proposed in this study is used to categorize three groups, as detailed in Section 4.1 . Ahi‘s algorithm was compared to the GT efficiency measure proposed by Chandrasekharan and Rajagopalan. The GT efficiency of the algorithm proposed in this study is 0.944, higher than the 0.887 shown in the literature for Ahi’s algorithm. Thus, the GT algorithm proposed in this study works and can perform better than some algorithms in the literature All parts in each group are to be operated on by at least one machine that is in the same group; this specification is applied using the proposed algorithm in this study. In Ahi’s algorithm, however, the two parts p 11 and p 14 belong to group G 2 , but all machines ( m 5 and m 6 ) that are in the same group do not operate on these two parts Example 2 contains four groups, as indicated in Section 4.2 . After discussions with the R&D engineers, however, it became clear that some parts, such as components p 5 , p 4 , and p 8 , could not be standardized due to their electrical characteristics, but some parts could be standardized. Part p 3 could be substituted for p 9 and p 11 . In addition, two new components were created: p 13 and p 14 , which replaced { p 2 , p 7 , p 10 } and { p 1 , p 12 , p 6 }, respectively. The specifications of p 13 and p 14 were 10.5 mm long AWG 18 and 9.5 mm long AWG 16, respectively The new sorting incidence matrix is as follows: Sustainability 2021 , 13 , x FOR PEER REVIEW 13 of 17 Example 2 contains four groups, as indicated in Section 4.2. After discussions with the R&D engineers, however, it became clear that some parts, such as components {?} , {?} , and {?} , could not be standardized due to their electrical characteristics, but some parts could be standardized. Part {?} could be substituted for {?} and {?} . In addition, two new components were created: {?} and {?} , which replaced { {?} , {?} , {?} } and { {?} , {?} , {?} }, respectively. The specifications of {?} and {?} were 10.5 mm long AWG 18 and 9.5 mm long AWG 16, respectively. {?} 1 3 {?} 5 {?} 1 4 {?} 3 {?} 4 {?} 8 {?} = {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ Example 3 features five groups. In group {?} , operators {?} and {?} are both present in the same group, which means they can perform similar tasks ( {?}{?} and {?}{?} ). If operator {?} takes leave, then operator {?} can be substituted for {?} immediately. In addition, if the manager wishes to train the operators on a limited budget, then they can choose certain operators to attend relevant courses (rather than sending all operators indiscriminately). The entry “0” is entered in the groups, as illustrated in the following sorted matrix. For instance, worker {?} can be trained to operate {?}{?} . Thus, the adverse effects of worker resignation or a sudden enforced quarantine due to the COVID-19 pandemic can be mitigated. {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} = {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ Example 3 features five groups. In group G 1 , operators o 10 and o 17 are both present in the same group, which means they can perform similar tasks ( wk 6 and wk 5 ). If operator o 10 takes leave, then operator o 17 can be substituted for o 10 immediately. In addition, if the manager wishes to train the operators on a limited budget, then they can choose certain operators to attend relevant courses (rather than sending all operators indiscriminately). The

Sustainability 2021 , 13 , 7880 14 of 17 entry “0” is entered in the groups, as illustrated in the following sorted matrix. For instance, worker o 9 can be trained to operate wk 13 . Thus, the adverse effects of worker resignation or a sudden enforced quarantine due to the COVID-19 pandemic can be mitigated Sustainability 2021 , 13 , x FOR PEER REVIEW 13 of 17 Example 2 contains four groups, as indicated in Section 4.2. After discussions with the R&D engineers, however, it became clear that some parts, such as components {?} , {?} , and {?} , could not be standardized due to their electrical characteristics, but some parts could be standardized. Part {?} could be substituted for {?} and {?} . In addition, two new components were created: {?} and {?} , which replaced { {?} , {?} , {?} } and { {?} , {?} , {?} }, respectively. The specifications of {?} and {?} were 10.5 mm long AWG 18 and 9.5 mm long AWG 16, respectively. The new sorting incidence matrix is as follows: {?} 1 3 {?} 5 {?} 1 4 {?} 3 {?} 4 {?} 8 {?} = {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} {?}{?}{?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ Example 3 features five groups. In group {?} , operators {?} and {?} are both present in the same group, which means they can perform similar tasks ( {?}{?} and {?}{?} ). If operator {?} takes leave, then operator {?} can be substituted for {?} immediately. In addition, if the manager wishes to train the operators on a limited budget, then they can choose certain operators to attend relevant courses (rather than sending all operators indiscriminately). The entry “0” is entered in the groups, as illustrated in the following sorted matrix. For instance, worker {?} can be trained to operate {?}{?} . Thus, the adverse effects of worker resignation or a sudden enforced quarantine due to the COVID-19 pandemic can be mitigated. {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} {?} = {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} {?}{?} ⎣ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎢ ⎡ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ⎦ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎥ ⎤ 6. Conclusions Rather than being lost to waste disposal, some used products and their components can be recycled. Using as few types of parts as possible in the manufacture of a product is ideal for two reasons: the business reaps the benefit of bulk buying and avoids risks of material shortages. Lacking just one type of material may result in product delivery delays, a suboptimal scenario for businesses. The fewer the types of parts employed, the lower is the business risk. One method of achieving this aim is component standardization, which can be used not only in manufacturing systems but also in remanufacturing systems In addition, all company managers want to reduce labor expenses and reduce the risk of sudden employee resignation. In this study, a sorting algorithm based on the concept of GT was proposed. There are limitations to the study’s approach: First, it assumes that the substitute worker has the same pay, performance, and workstation efficiency as the original worker; second, it assumes that the substitute component does not affect the product’s performance. This algorithm facilitates component standardization and worker rotation One numerical example and two case studies were presented to demonstrate the utility of the proposed algorithm. The first example demonstrated that the performance of the algorithm proposed in this study is superior to that of a previously described algorithm The second example revealed that an original range of 12 distinct parts can be reduced to 6 by using the algorithm proposed in this work. Making provisions for emergency business scenarios is crucial; thus, the final case study demonstrated how to group workers with similar operational abilities and suggested how to assign tasks to alternate workers if a worker resigns suddenly or is quarantined due to the pandemic alleviation measures Traditional GT considers only machines and parts, but this study uses the concept of GT to solve the issues of component standardization and absent workers. The proposed method can reduce the number of parts and mitigate the risk of an unexpected worker absence. As this is the era of AMS, different operation processes, robots, and operators must all be considered. Future studies could examine how companies can use algorithms to manage robots breaking down or requiring maintenance at the same time that workers are absent Funding: This research received no external funding Institutional Review Board Statement: Not applicable Informed Consent Statement: Not applicable Data Availability Statement: Not applicable.

Sustainability 2021 , 13 , 7880 15 of 17 Conflicts of Interest: The author declares no conflict of interest Nomenclature m Number of machines n Number of parts A = n a ij o m × n Incidence matrix m i ith machine p i ith part OS = n os ij o m × n Operational sequences of the incidence matrix os ij os ij th operating process of part j SC m Similarity coefficient matrix of machines sc ij Similarity coefficient of m i and m j W Maximum similarity coefficient matrix O Ordering matrix Λ Association set matrix MC m Number of machine cells MC i ith machine cell PF n Number of part families PF i ith part family A T Transpose of incidence matrix A G i ith group η Group efficiency e b Total number of 1 s in the major blocks e 0 Total number of 1 s in the stray blocks Q i Number of machines in the ith cell P i Number of parts in the ith family q Weighting factor pdt i ith product or semiproduct wk i ith workstation References 1 LaMarco, N. Types of Manufacturing Systems. Available online: https://bizfluent.com/info-7899360-types-manufacturingsystems.html (accessed on 22 June 2021) 2 Esmaeilian, B.; Behdad, S.; Wang, B. The evolution and future of manufacturing: A review J. Manuf. Syst 2016 , 39 , 79–100 [ CrossRef ] 3 Gola, A. Design and Management of Manufacturing Systems Appl. Sci 2021 , 11 , 2216. [ CrossRef ] 4 Vidal, G.H.; Coronado Hern á ndez, J.R. Complexity in manufacturing systems: A literature review Prod. Eng 2021 , 15 , 321–333 [ CrossRef ] 5 Fratini, L.; Ragai, I.; Wang, L. New trends in Manufacturing Systems Research 2020 J. Manuf. Syst 2020 , 56 , 585–586. [ CrossRef ] [ PubMed ] 6 Mital, A.; Pennathur, A.; Huston, R.; Thompson, D.; Pittman, M.; Markle, G.; Kaber, D.; Crumpton, L.; Bishu, R.; Rajurkar, K.; et al The need for worker training in advanced manufacturing technology (AMT) environments: A white paper Int. J. Ind. Ergon 1999 , 24 , 173–184. [ CrossRef ] 7 Ke, Q.; Zhang, H.-C.; Liu, G.; Li, B. Remanufacturing Engineering Literature Overview and Future Research Needs. In Glocalized Solutions for Sustainability in Manufacturing ; Hesselbach, J., Herrmann, C., Eds.; Springer: Berlin/Heidelberg, Germany, 2011; pp. 437–442. ISBN 978-364-219-691-1 8 Aljuneidi, T.; Bulgak, A.A. Designing a Cellular Manufacturing System featuring remanufacturing, recycling, and disposal options: A mathematical modeling approach CIRP J. Manuf. Sci. Technol 2017 , 19 , 25–35. [ CrossRef ] 9 Xiaoyan, W. Research on design management based on green remanufacturing engineering Syst. Eng. Procedia 2012 , 4 , 448–454 [ CrossRef ] 10 Cao, J.; Zhang, X.; Hu, L.; Xu, J.; Zhao, Y.; Zhou, G.; Schnoor, J.L. EPR regulation and reverse supply chain strategy on remanufacturing Comput. Ind. Eng 2018 , 125 , 279–297. [ CrossRef ] 11 Mohamed, B.; Rami, A.; Kanwar, A.N. A single-vendor multi-buyer production remanufacturing inventory system under a centralized consignment arrangement Comput. Ind. Eng 2019 , 135 , 10–27. [ CrossRef ] 12 Dias, A.S.; Kim, H.; Sivakumar, P.K.; Liu, Z.C.; Zhang, H.C. Life Cycle Assessment: A Comparison of Manufacturing and Remanufacturing Processes of a Diesel Engine. In Re-Engineering Manufacturing for Sustainability ; Nee, A., Song, B., Ong, S.K., Eds.; Springer: Singapore, 2013; pp. 675–678.

Sustainability 2021 , 13 , 7880 16 of 17 13 Milios, L.; Beqiri, B.; Whalen, K.A.; Jelonek, S.H. Sailing towards a circular economy: Conditions for increased reuse and remanufacturing in the Scandinavian maritime sector J. Clean. Prod 2019 , 225 , 227–235. [ CrossRef ] 14 Xu, B.-S.; Liu, S.-C.; Wang, H.-D. Developing remanufacturing engineering, constructing cycle economy and building savingoriented society J. Cent. South Univ. Technol 2005 , 12 , 1–6. [ CrossRef ] 15 Hendrickson, C.; Matthews, D.H.; Ashe, M.; Jaramillo, P.; McMichael, F.C. Reducing environmental burdens of solid-state lighting through end-of-life design Environ. Res. Lett 2010 , 5 , 014016. [ CrossRef ] 16 Das, D.; Dutta, P. A system dynamics framework for integrated reverse supply chain with three way recovery and product exchange policy Comput. Ind. Eng 2013 , 66 , 720–733. [ CrossRef ] 17 Xu, B.-S. Progress of remanufacturing engineering and future technology expectation Adv. Manuf 2013 , 1 , 8–12. [ CrossRef ] 18 Clarke, J.W.; White, M.S.; Araman, P.A. Comparative performance of new, repaired, and remanufactured 48-by 40-inch GMA-style wood pallets For. Prod. J 2005 , 55 , 83–88 19 Shu, T.; Xu, J.; Chen, S.; Wang, S.; Lai, K.K. Remanufacturing Decisions with WTP Discrepancy and Uncertain Quality of Product Returns Sustainability 2018 , 10 , 2123. [ CrossRef ] 20 Fatimah, Y.A.; Govindan, K.; Murniningsih, R.; Setiawan, A. Industry 4.0 based sustainable circular economy approach for smart waste management system to achieve sustainable development goals: A case study of Indonesia J. Clean. Prod 2020 , 269 , 122263 [ CrossRef ] 21 Yan, G.; Ni, Y.; Yang, X. Optimal Pricing in Recycling and Remanufacturing in Uncertain Environments Sustainability 2020 , 12 , 3199. [ CrossRef ] 22 Perera, H.S.C.; Nagarur, N.; Tabucanon, M.T. Component part standardization: A way to reduce the life-cycle costs of products Int. J. Prod. Econ 1999 , 60–61 , 109–116. [ CrossRef ] 23 Ko, Y.D.; Noh, I.; Hwang, H. Cost benefits from standardization of the packaging glass bottles Comput. Ind. Eng 2012 , 62 , 693–702 [ CrossRef ] 24 Standardization in Material Management. Available online: https://www.wisdomjobs.com/e-university/production-andoperations-management-tutorial-295/standardization-9600.html (accessed on 22 June 2021) 25 Szwarc, E.; Wikarek, J. Proactive Planning of Project Team Members’ Competences Found. Manag 2020 , 12 , 71–84. [ CrossRef ] 26 Szwarc, E.; Bocewicz, G.; Bach-D ˛ abrowska, I.; Banaszak, Z. Declarative Model of Competences Assessment Robust to Personnel Absence. In Proceedings of the International Conference on Information and Software Technologies (ICIST 2019), Vilnius, Lithuania, 10–12 October 2019; pp. 12–23 27 Szwarc, E.; Wikarek, J.; Gola, A.; Bocewicz, G.; Banaszak, Z. Interactive Planning of Competency-Driven University Teaching Staff Allocation Appl. Sci 2020 , 10 , 4894. [ CrossRef ] 28 Agard, B. Standardization of Components, Products and Processes with Data Mining. Available online: https://docplayer.net/14 179790-Standardization-of-components-products-and-processes-with-data-mining.html (accessed on 22 June 2021) 29 Classification and Part Standardization. Available online: http://enfinio.com/classification-and-part-standardization/ (accessed on 22 June 2021) 30 Anderson, D.M. Part Standardization. Available online: http://www.build-to-order-consulting.com/Standardization.htm (accessed on 22 June 2021) 31 Wemmerlöv, U.; Hyer, N.L. Cellular manufacturing in the U.S. industry: A survey of users Int. J. Prod. Res 1989 , 27 , 1511–1530 [ CrossRef ] 32 Hyer, N.; Wemmerlov, U Reorganizing the Factory ; Productivity Press: Portland, OR, USA, 2002; p. 48. ISBN 1-56327-228-8 33 Hazarika, M.; Laha, D. Genetic Algorithm approach for Machine Cell Formation with Alternative Routings Mater. Today Proc 2018 , 5 , 1766–1775. [ CrossRef ] 34 Mohammadi, M.; Forghani, K. Designing cellular manufacturing systems considering S-shaped layout Comput. Ind. Eng 2016 , 98 , 221–236. [ CrossRef ] 35 Shahdi-Pashaki, S.; Teymourian, E.; Kayvanfar, V.; Komaki, G.; Sajadi, A. Group technology-based model and cuckoo optimization algorithm for resource allocation in cloud computing IFAC-PapersOnLine 2015 , 48 , 1140–1145. [ CrossRef ] 36 Zohrevand, A.M.; Rafiei, H.; Zohrevand, A.H. Multi-objective dynamic cell formation problem: A stochastic programming approach Comput. Ind. Eng 2016 , 98 , 323–332. [ CrossRef ] 37 Liu, C.; Wang, J.; Leung, J.Y.-T. Integrated bacteria foraging algorithm for cellular manufacturing in supply chain considering facility transfer and production planning Appl. Soft Comput 2018 , 62 , 602–618. [ CrossRef ] 38 Nalluri, M.S.R.; Kannan, K.; Gao, X.-Z.; Roy, D.S. An efficient hybrid meta-heuristic approach for cell formation problem Soft Comput 2019 , 23 , 9189–9213. [ CrossRef ] 39 Li, M.-L. Arrangement of Machines Under Spatial Constraint by Using a Novel Algorithm IEEE Access 2020 , 8 , 144565–144574 [ CrossRef ] 40 Padayachee, J.; Bright, G. A method for optimising the evolution of cellular manufacturing systems. In Proceedings of the 2017 24 th International Conference on Mechatronics and Machine Vision in Practice (M 2 VIP), Auckland, New Zealand, 1–6 November 2017; pp. 1–6 41 Pimentel, C.; Martins, S. Design and implementation of a manufacturing cell in a job shop environment: An action research study. In Proceedings of the 2015 International Conference on Industrial Engineering and Systems Management (IESM 2015), Seville, Spain, 21–23 October 2015; pp. 398–407.

Sustainability 2021 , 13 , 7880 17 of 17 42 Miao, C. Parallel-Batch Scheduling with Deterioration and Group Technology IEEE Access 2019 , 7 , 119082–119086. [ CrossRef ] 43 Boe, W.J.; Cheng, C.H. A close neighbour algorithm for designing cellular manufacturing systems Int. J. Prod. Res 1991 , 29 , 2097–2116. [ CrossRef ] 44 Rkin, R.E.P.; Li, M.-L. The multi-dimensional aspects of a group technology algorithm Int. J. Prod. Res 1997 , 35 , 2345–2358 [ CrossRef ] 45 Gupta, N.S.; Devika, D.; Valarmathi, B.; Sowmiya, N.; Shinde, A. CARI—A heuristic approach to machine-part cell formation using correlation analysis and relevance index Int. J. Prod. Res 2014 , 52 , 7297–7309. [ CrossRef ] 46 Danilovic, M.; Ilic, O. A novel hybrid algorithm for manufacturing cell formation problem Expert Syst. Appl 2019 , 135 , 327–350 [ CrossRef ] 47 Won, Y. On Solving Large-Size Generalized Cell Formation Problems via a Hard Computing Approach Using the PMP Appl. Sci 2020 , 10 , 3478. [ CrossRef ] 48 Won, Y.; Logendran, R. Effective two-phase p-median approach for the balanced cell formation in the design of cellular manufacturing system Int. J. Prod. Res 2014 , 53 , 2730–2750. [ CrossRef ] 49 Rahimi, V.; Arkat, J.; Farughi, H. A vibration damping optimization algorithm for the integrated problem of cell formation, cellular scheduling, and intercellular layout Comput. Ind. Eng 2020 , 143 , 106439. [ CrossRef ] 50 McAuley, J. Machine grouping for efficient production Prod. Eng 1972 , 51 , 53–57. [ CrossRef ] 51 Suzic, N.; Stevanov, B.; ´ Cosi´c, I.; Aniši´c, Z.; Sremˇcev, N. Customizing Products through Application of Group Technology: A Case Study of Furniture Manufacturing J. Mech. Eng 2012 , 58 , 724–731. [ CrossRef ] 52 Kumar, R.; Singh, S.P. A similarity score-based two-phase heuristic approach to solve the dynamic cellular facility layout for manufacturing systems Eng. Optim 2017 , 49 , 1848–1867. [ CrossRef ] 53 Rajesh, K.D.; Krishna, M.M.; Ali, A.; Chalapathi, P. A Modified Hybrid Similarity Coefficient Based Method for Solving the Cell Formation Problem in Cellular Manufacturing System Mater. Today Proc 2017 , 4 , 1469–1477. [ CrossRef ] 54 Martins, I.C.; Pinheiro, R.G.S.; Protti, F.; Ochi, L.S. A hybrid iterated local search and variable neighborhood descent heuristic applied to the cell formation problem Expert Syst. Appl 2015 , 42 , 8947–8955. [ CrossRef ] 55 Laha, D.; Hazarika, M. A heuristic approach based on Euclidean distance matrix for the machine-part cell formation problem Mater. Today Proc 2017 , 4 , 1442–1451. [ CrossRef ] 56 Feng, Y.; Li, G.; Sethi, S.P. A three-layer chromosome genetic algorithm for multi-cell scheduling with flexible routes and machine sharing Int. J. Prod. Econ 2018 , 196 , 269–283. [ CrossRef ] 57 Hazarika, M.; Laha, D. Machine-part cell formation for maximum grouping efficacy based on genetic algorithm. In Proceedings of the 2015 IEEE Workshop on Computational Intelligence: Theories, Applications and Future Directions (WCI), Kanpur, India, 14–17 December 2015; pp. 1–6 58 Imran, M.; Kang, C.; Lee, Y.H.; Jahanzaib, M.; Aziz, M.H. Cell formation in a cellular manufacturing system using simulation integrated hybrid genetic algorithm Comput. Ind. Eng 2017 , 105 , 123–135. [ CrossRef ] 59 Hazarika, M.; Laha, D. Application of genetic algorithm in generalized machine cell formation problem. In Proceedings of the 2017 International Conference on Energy, Communication, Data Analytics and Soft Computing (ICECDS), Chennai, India, 1–2 August 2017; pp. 1855–1860 60 Yu, F.; Tu, F.; Pattipati, K. A novel congruent organizational design methodology using group technology and a nested genetic algorithm IEEE Trans. Syst. Man Cybern. Part A Syst. Hum 2005 , 36 , 5–18. [ CrossRef ] 61 Li, M.-L. The algorithm for integrating all incidence matrices in multi-dimensional group technology Int. J. Prod. Econ 2003 , 86 , 121–131. [ CrossRef ] 62 Li, M.-L. A novel algorithm of cell formation with alternative machines and multiple-operation-type machines Comput. Ind. Eng 2021 , 154 , 107172. [ CrossRef ] 63 Huang, S.; Yan, Y. Part family grouping method for reconfigurable manufacturing system considering process time and capacity demand Flex. Serv. Manuf. J 2018 , 31 , 424–445. [ CrossRef ] 64 Seitz, M.A. A critical assessment of motives for product recovery: The case of engine remanufacturing J. Clean. Prod 2007 , 15 , 1147–1157. [ CrossRef ] 65 Chandrasekharan, M.P.; Rajagopalan, R. MODROC: An extension of rank order clustering for group technology Int. J. Prod. Res 1986 , 24 , 1221–1233. [ CrossRef ] 66 Ahi, A.; Aryanezhad, M.; Ashtiani, B.; Makui, A. A novel approach to determine cell formation, intracellular machine layout and cell layout in the CMS problem based on TOPSIS method Comput. Oper. Res 2009 , 36 , 1478–1496. [ CrossRef ] 67 Boulif, M.; Atif, K. A new branch-&-bound-enhanced genetic algorithm for the manufacturing cell formation problem Comput Oper. Res 2006 , 33 , 2219–2245. [ CrossRef ]