Growing environmental concerns and social legislations enforce decision makers to design their supply chains considering environmental and social impacts as well as economical objectives. Degradation difficulties and recovering profits lead to recycle scraped tires regarding the sustainability factors. This paper firstly develops a multi-objective mixed integer linear programming model for designing of sustainable tire closed-loop supply chain network. The proposed model aims to optimize total cost, environmental impacts of establishment of facilities, processing of tires and transportation between each level as well as social impacts including job opportunities and work's damages. To alleviate the drawbacks of existing metaheuristic algorithms when solving the large-scale networks, four new hybrid metaheuristic algorithms based on the advantages of recent and old ones are developed. To evaluate the quality of the proposed hybrid algorithms, extensive computational experiments, comparison, and sensitivity analyses are conducted with different criteria. Results reveal that hybrid algorithms are effective approaches to solve the underlying problem in large-scale networks.
According to the U.S. Environmental Protection Agency (EPA) report, world demand for tires is increasing 4.1 percent per year and reaches to 3.0 billion units in 2019 (Amin et al., 2017). Hence, economic design of tire supply chain network became an important issue for both academics and practitioners (Stadtler, 2015). In addition, EPA reported that about 290 million scraped tires are disposed every year. Unfortunately, the amount of the scraped tires that are released in the nature without environmental consideration threaten human health, water, air, and soil condition (Subulan et al., 2015a). Concerns aroused when twenty percent of scarped tires are illegally dumped in landfills or roadsides. To tackle these issues, supply chain managers need to develop sustainable network for managing tire supply chain considering economic, social, and environmental factors (Pishvaee et al., 2014). Whereas, in the literature less attentions has been paid for designing sustainable tire closed-loop supply chain network (Govindan et al., 2016).
Todays, growth of recycling technologies and environmental regulations led to convert recyclable items into raw material that can be used in new products (Amin et al., 2017). Dehghanian and Mansour (2009) designed recovery network to mitigate the negative environmental and social impact of the end-of-life products such as tire. They used Life-Cycle Assessment (LCA) based methodology for estimating social and environmental impacts and indicated different measures for each stage. Identifying the social and environment measures, the stages of the product's life cycle, and estimating the overall impact of products were presented as main challenges of the LCA-based methodology (Dehghanian and Mansour, 2009). While the authors aim to integrate the forward and reverse flows of supply chains to improve the performance of the networks, Dehghanian and Mansour (2009) considers forward flow between collection centers and recycling plants. Therefore authors aims to develop sustainable supply chains by considering forward and revers flows. For example, Diabat et al. (2015) considered remanufacturing centers in supply chain that were collecting recyclable items, remanufacturing them, and then distribute them among retailer. Kaya and Urek (2016) designed closed-loop supply chain network considering both distribution and collection decisions. To increase the total profit of supply chain network, they determined incentive values for collecting right amount of recyclable items. Ozceylan et al. (2016) € designed closedloop supply chain network for recycling end-of-life vehicles under regulation by Republic of Turkey Ministry of Environment and Urbanization. Zhalechian et al. (2016) presented sustainable closedloop supply chain model by considering CO2 emissions, fuel and energy consumption, and the social impacts of new job opportunities. Recently, Soleimani et al. (2017) proposed three kinds of recycling i.e. product, components, and raw material by accounting environmental impact, total profit of network, and lost working. Opening or closing of network's node and product flow among them were determined by concerning responsiveness to customer demand as well as sustainability factors. Among studies have modeled the closed-loop supply chain problems; there still exists a gap in the modeling of economic, environmental and social dimensions concurrently (Samadi et al., 2018).
To cope with the described problem, this study designed sustainable tire closed-loop supply chain network considering the economic, environmental and social dimensions. Different measures for each stage of manufacturing, recycling, and transportation along supply chain has been presented to estimate the social and environmental impact of tire in closed-loop supply chain. Four new hybrid multi-objective metaheuristic algorithms were developed to encounter with the complexity of the large scale networks. The proposed algorithms are tested with different random generated test problems. In addition, to generalize the proposed algorithms, ten standard benchmarked functions were solved and results are compared with well-known multi-objective algorithms. The main contributions of this paper can be outlined as follows:
Developing a new model of the sustainable tire closed-loop supply chain network for the first time;
Proposing four new hybrid meta-heuristic algorithms to reduce the computational time and improve the quality of solutions;
Generalizing the performance of the proposed hybrid metaheuristic algorithms using standard benchmarked problems;
Developing the LCA-methodology for estimating the sustainability dimensions of the tire closed-loop supply chain network.
The rest of the paper is organized as follows. The relevant literature is reviewed in Section 2. The proposed model of tire sustainable closed-loop supply chain network is formulated in Section 3. Section 4 presents the hybrid meta-heuristic algorithms along with brief review of basic algorithms. Computational experiments, comparison, and sensitivity analyses are reported in section 5. Finally, Section 6 provides concluding remarks and directions for further research.
2. Literature review
Generally, supply chain management aims to find the best strategies for controlling and managing the supply chains. One of the well-known problems in supply chain management is Supply Chain Network Design (SCND) (Eskandarpour et al., 2017). Scholars solve SCND problems for improving the products flow between different levels such as suppliers, manufacturers, retailers, and customers (Badri et al., 2013). Several researches in the relevant literature have focused on SCND problems considering different features of real cases such as products' characteristics (Coelho and Laporte, 2014), number of levels (Srivathsan and Kamath, 2017), network flows (Ivanov et al., 2017), transportation (Cui et al., 2016), effects of different types of technologies (Fard et al., 2017), specifications of supplier, manufacturer, and distributer facilities (Shafiee et al., 2014). Indeed, the SCND problems become more complicated considering more features of real cases. To design tire supply chain network different features should be considered including
i) tires are recyclable item and should be recycled and remanufactured;
ii) a network should be designed with four levels consisting of raw material suppliers, manufacturer/recycler, retailer/collection center, customer;
iii) forward flows start from supplier and end to customer and reverse flow from collection center to manufacturer or supplier. Taken together, the scholars have developed the closed-loop supply chain for designing network of recyclable items along with considering the sustainable dimensions (Paydar et al., 2017).
2.1. Closed-loop supply chain
Regarding the recent recycling and remanufacturing technologies developed, the researchers have paid more attentions to integrate forward and reverse logistics as closed-loop supply chain network (Xie et al., 2017). Among them, Ozceylan et al. (2016) € emphasized that environmental and social concerns and legislations motivate the decision makers for designing closed-loop supply chains. A linear programming model has been developed to distribute new vehicles and collect end-of-life vehicles from customers. However their model was not considered the facility locations decisions and multiple suppliers. In addition, they suggested that heuristics approaches should be developed to cope with large scale networks. Mohammed et al. (2017) optimized industry environmental impact and total cost for designing and planning a multi-period, multi-product closed-loop supply chain network. Their model evaluated the trade-offs between carbon emissions as industry environmental impact and total cost. Although their multi-stage scenario-based stochastic approaches could provide robust solutions, for solving real cases (i.e. large-scale networks) desirable approaches should be developed (Mohammed et al., 2017). Paydar et al. (2017) presented a bi-objective optimization model for closed-loop supply chain of used engine oil. Their model maximizing profit and minimizing risk of closed-loop supply chain under different scenarios. Increasing the size of problems, their solution approach is encountering with computational cost. Most of scholars confirmed that the complexity of underling problem would be increased significantly to deal with large scale networks (Fahimnia et al., 2015). This reason motivated several researchers like this study to contribute the new capable metaheuristic algorithms for solving such problems.
2.2. Sustainable supply chain
network design Another issue which should be considered for developing closed-loop supply chain network is to consider the sustainability dimensions. Regarding the real applications of this issue, recent studies considered different types of objective functions to develop the sustainable supply chain network design (Soleimani et al., 2013). Note that most of researches focus on economic objective function such as profit or cost. Although more attentions have been paid on the environmental impact of supply chain recently, limit researches were considered the social impact of closed-loop supply chain yet. Scholars should be aimed to integrate the environmental impacts of supply chains' operations as well as the health and safety of employees and society (Zhalechian et al., 2016). Sustainability theory leads decision makers to integrate economic, environmental, and social dimensions for designing closed-loop supply chain network (Boukherroub et al., 2015; Genovese et al., 2017). Mota et al. (2015) developed a multi-objective mathematical programming model for planning closed-loop supply chain. LCA as a reliable method for evaluating environmental impact (introduced by European CommissionE, 2010), negative impact of unemployed people in society, and total cost of network were considered for designing network. They are emphasized that social factors should be improved for evaluating social dimension of supply chain.proposed a multi-objective mathematical model aim to optimize CO2 emissions, profit of the chain, and the number of missed working days. They highlighted that more attempt need for quantifying the social and environmental impact as well as meta-heuristic algorithms to solve large-scale networks. Nevertheless few studies have been developed quantitative approaches for adopting the whole dimensions of the sustainability in the closed-loop supply chain. Recently, Ansari and Kant (2017) explored the main knowledge gaps and research opportunities of sustainable supply chain management. Quantifying the social and environmental dimensions of the sustainability, implementing sustainability in real complex cases such as automobile industries, and multi-objective meta-heuristic algorithms were introduced as main gaps in the sustainable closed-loop supply chain literature (Ansari and Kant, 2017).