Sustainable supply chain network design problem: Using the integrated BWM, TOPSIS, possibilistic programming, and ε-constrained methods

Highlights

• Combined best-worst, ε-constraint, and possibilistic programming is used.

• Supplier's green image factors identified and used in the mathematical model.

• Epsilon constraint method is used to generates distinct Pareto-optimal solutions.

Abstract

This paper proposes a two-stage multi-objective possibilistic integer linear programming sustainable supply chain network design model, minimizing the economic, environmental goals and maximizing the social sustainability goals. The proposed model determines the openings of facilities and the amount of flow of goods across the supply chain. It introduces supplier green image factors in the design of the supply chain network. The model has considered epistemic uncertainty to model the unknown capacity, cost, and demand. The proposed study has been carried out in two stages. In the first stage, BWM (Best-Worst method) and TOPSIS are applied to evaluate the green image weights of suppliers. Further, these green weights are being used in the second phase for the supply chain network design. The study has adopted combined possibilistic programming and Epsilon (ε) constraint method, which was reported infrequently in the literature. Epsilon (ε) constraint method generates distinct Pareto-optimal solutions, which provided a large combination of the trade-off between the cost, emission, and social sustainability. The results facilitate decision-makers to take the decision in an uncertain environment.

Introduction

Supply chain network design (SCND) is the elaboration of a complex structure of interaction between organizations for managing the flow of material, information, and money. In the early 90s, the prime concern of SCND was to maximize the overall profit and/or minimizing the costs by optimizing the material flow across the SC (Farahani et al., 2014, Pishvaee and Torabi, 2010). Due to diversifying consumer decisions towards greener products, rising regulatory pressures from the governments, and Non-Governmental Organizations, companies are now focusing more on the design of their SC to make them sustainable (de Oliveira et al., 2018, Wu and Pagell, 2011). The elucidation of sustainability was given in the Brundtland report as “meeting the needs of the present generations without compromising the needs of future generations” (Keeble, 1988). The sustainability concept often addresses the three major dimensions: social, economic, and environmental. Recently, sustainable supply chain network design (SSCND) has received notable consideration from scholars. Guo, Hu, Allaoui, and Boulaksil (2019) observed SSCND as a challenging task. Despite its importance, very limited studies have been conducted, especially addressing three dimensions of sustainability. In archived literature, only 17 percent of SCND models addressed sustainability issues (Govindan, Fattahi, & Keyvanshokooh, 2017). Hence, there is enormous scope for developing quantitative models for SSCND (Gunasekaran & Spalanzani, 2012).

SCND is a long-term strategic decision. Due to its complexity, it cannot be easily changed once it is formed (Tsao, Thanh, Lu, & Yu, 2018). The SCND is indeed an important part in implementing sustainability, as conventional SC activities account for approximately 50–70% of carbon emissions (Ahmed & Sarkar, 2018). Waltho, Elhedhli, and Gzara (2019) discussed different greenhouse gas reduction plans, such as carbon offset, carbon cap, cap-and-trade, and carbon tax, to lower the emission across the SC. However, limited studies were conducted addressing uncertainty and social sustainability. Social sustainability is defined as “developing a sustainable environment which upgrades living standards, social wellbeing by taking care of the need of the people where they live and work (McKenzie, 2004). In business organizations, social sustainability can be measured through labor working conditions, health and safety, human rights, fair labor wages, employment generation, work-life balance, equality of genders, empowerment, and more. Quantifying social sustainability is a difficult job, which needs to be addressed in the SCND (Das and Shaw, 2017, Tsao et al., 2018).

In addition to the sustainability dimension, a manager has to make decisions in a dynamic environment. A dynamic environment inflicts a massive level of uncertainty in decision making (Pishvaee & Torabi, 2010). Managing uncertainty is a real challenge for business. Currently, SSCND models have been formulated, taking deliberation of a deterministic environment, and limited models have been reported in a stochastic environment. To overcome these gaps, this study proposes a two-stage multi-objective possibilistic integer linear programming (MPILP) model, considering the three sustainability dimensions. As a theoretical contribution, we introduce a new sustainability dimension, supplier’s green image (SGI), in the design phase of SCND. In the first stage, SGI has been quantified by applying BWM and TOPSIS methods. The ratings have been subsequently integrated into the mathematical model.

In the second stage, a fuzzy possibilistic technique is proposed to handle uncertain scenarios. Further, this study applies the ε-constraint method for getting trade-off relationships among the three conflicting objectives. This work aims to reply to the following research questions:

• Which facilities should be opened to minimize the emission and cost of the supply chain?

• What would be the amount of material flow across the different echelons of the supply chain in an uncertain situation?

• What are the impacts of changing the overall carbon cap on the movements of goods across the chain?

• What are the effects of uncertainty over the model?

• What are the computational performances of the model in different scenarios?

The paper is organized as follows. The appropriate literature is presented in Section 2. Section 3 comprises the model development. The proposed methodology is presented in Section 4. Computational experiments are stated in Section 5. Lastly, conclusion, the managerial implication of the study and future scope are drawn in Section 6.