Taguchi Analysis for Improving Optimization of Integrated Forward/Reverse Logistics

Journal of the Operations Research Society of China https://doi.org/10.1007/s40305-021-00380-7 Taguchi Analysis for Improving Optimization of Integrated Forward/Reverse Logistics Elham Behmanesh1 · Jürgen Pannek2,3 Received: 24 May 2019 / Revised: 4 October 2021 / Accepted: 2 November 2021 © The Author(s) 2022 Abstract The distribution–allocation problem is known as one of the most comprehensive strategic decisions. In real-world cases, it is impossible to solve a distribution–allocation problem completely in acceptable time. This forces the researchers to develop efficient heuristic techniques for the large-term operation of the whole supply chain. These techniques provide near optimal solution and are comparably fast particularly for large-scale test problems. This paper presents an integrated supply chain model which is flexible in the delivery path. As solution methodology, we apply a memetic algorithm with a novelty in population presentation. To identify the optimum operating condition of the proposed memetic algorithm, Taguchi method is adopted. In this study, four factors, namely population size, crossover rate, local search iteration and number of iteration, are considered. Determining the best level of the considered parameters is the outlook of this research. Keywords Integrated logistics network · Flexible path · Memetic algorithm · Taguchi analysis Mathematics Subject Classification 90B06 · 68W50 · 91G70 1 Introduction Supply chain networks describe the flow and movement of goods by linking several facilities such as plants, distributions, and retailers in forward flow[1,2] and collec- B Elham Behmanesh Jürgen Pannek 1 International Graduate School for Dynamics in Logistics, University of Bremen, Bremen, Germany 2 Faculty of Production Engineering, University of Bremen, Bremen, Germany 3 BIBA Bremer Institut für Produktion und Logistik GmbH, Bremen, Germany 123 E. Behmanesh, J. Pannek tion/inspection centers and disposal centers in backward flow [3]. Supply chain goals are summarized in minimization of total cost [4,5] or maximization of profit [6] according to the interested outcome. Supply chain activities involve determining the optimal number and capacity of facilities as well as the flow between them [5]. Network configuration comes first in any supply chain network design and needs to be optimized for a long-lasting efficient operation of the entire supply chain. Therefore, long-term strategic decisions take priority over tactical and operational levels. A result of a comprehensive review of the supply chain literature by Thomas and Griffin [7] shows that distribution cost expenditures are equal to about 30 percent of product costs. Hence, focusing on decreasing transportation cost has a significant impact on reducing the distribution cost of the supply chain network. In this regard, considering flexibility in delivery path to have optional short ways delivery is noticeable. Besides, fast and on-time delivery of products plays an important role in customer satisfaction [8,9]. Although different delivery paths improve efficiency, it reveals the problem more complex. But the problem is, the purpose of reducing delivery time is often conflict with the goal of reducing logistics cost [10]. To deal with the issue of cost efficiency and network responsiveness, researchers have proposed models to optimize both but results are typically limited to shipments between consecutive stages or just indirect shipment mechanisms [8,9,11]. The full capacitated graph in forward flow considered in this study allows us to solve conflicting goals of profit and responsiveness which otherwise may lead to greater cost. As environmental protection forced firms to pay more attention to collect, recover, recycle and safe disposal in a supply chain network [12], reverse distribution needs to be added in any supply chain network. While the reverse activity could be useful for environmental protection, industries can use returned product for economic benefits [13]. Although industrial players are forced to handle returned products, most of logistics networks are not equipped to deal with this requirement [12]. Therefore, management of product return flow is becoming an essential part of each supply chain. Within this paper, we attempt to include the reverse flow through an integrated design of forward/reverse supply chain network. The proposed integrated design leads us to a closed-loop supply chain network which is avoid sub-optimal solutions derived by separated design [14], cf. Fig. 1 for a sketch. As shown in Fig. 1, we have raw materials from supplier to plant. New products are shipped from plant to customer through distribution center and retailer. Three different delivery paths are considered in the forward flow to enrich the model in order to being close to customer, reducing transportation cost and increasing customer satisfaction. Except normal delivery, which is from any stage to another adjoining one, we added direct shipment and direct delivery. In direct delivery, goods are transported from distribution centers to customers by skipping retailers or from plants to retailers by skipping distribution centers. In direct shipment, products are transferred from plants to customers directly. In the backward flow, returned products are collected by collection centers and, after inspection, the recoverable products are shipped to plants, and scraped products are transferred to disposal centers for a safe disposal. According to the aforementioned description, in this study, we attempt to add the reverse flow through an integrated design to the presented network. Also, a full delivery graph in forward flow between plants and customers is considered to increase the performance of the supply chain network. This model can be formulated into an 123 Taguchi Analysis for Improving Optimization of Integrated· · · Fig. 1 Framework of the proposed closed-loop supply chain network integer linear programming, which traditional methods fail to solve in acceptable time, particularly when we are facing with large size problems. To overcome this problem, nontraditional solutions such as heuristics and metaheuristics [8,11,15–17] have been proposed in recent years. Within this paper, we present a memetic algorithm with a new chromosome representation as well as updated operators. Each algorithm has some parameters. Some information regarding these parameters would be useful to improve the performance of the results. In this study, Taguchi method is applied to improve operation condition. Determining the best level of parameters is the main contribution of this work. In the last several decades, growing attention is being paid to the use of Taguchi method to find the best setting of parameters involved in evolutionary algorithms. Chouhan et al. [18] introduced a novel approach that had the ability in collecting endof-life and end-of-use products from the end-users and i (...truncated)