Multi-tier supply chain network design: A key towards sustainability and resilience

Abstract

The inescapable need for sustainable supply chains has fascinated controversial attention, in the last two decades, due to increasing governmental regulations and societal awareness. But, facilities and their linkages, within the supply chain network, are subject to perturbations due to natural or human-made events (e.g., strikes and floods). Therefore, managers are concerned about re-engineering a supply chain network that is simultaneously sustainable and resilient. Hence, there is an inevitable need to embrace sustainability and resilience at a strategic level for the supply chain management. Motivated by this necessity, this paper presents a new panacea methodology towards a sustainable and resilient (susilient, henceforth) two-tier supply chain network design (S-2TSCND). To this end, an extended exploration regarding the relationship and completeness between sustainability and resilience, in the supply chain context, was explored and discussed. It presents a new term for “susilient” supply chain due to the strong correlation found between these two paradigms. Then, a holistic framework was developed to identify dimensions, enablers, and criteria towards susilient development. Next, a fuzzy four-objective optimization model (FFOOM) was developed to (1) solve the susilient facility location problem; (2) assign the optimal order quantities among the selected facilities considering the susilience aspect; and (3) derive a trade-off between sustainability and resilience. The susilience values of nominated facilities were quantified via a hybrid AHP-OCRA method. These values were then merged into the FFOOM to consolidate susilience performance into the strategic and tactical decisions. A set of trade-offs was then derived via the ε-constraints method. Finally, each trade-off was examined via the global criterion approach to measure its distance from the ideal solution aiming to select the final supply chain network design. The validity of the developed methodology was proved on a re-configuration of an existing real 2TSCND.