A two-stage stochastic optimization model for reverse logistics network design under dynamic suppliers' locations

Abstract

This paper presents a two-stage stochastic programming model for reverse logistics network design (RLND) under uncertainty with dynamic supply sources' locations. The primary objective of the optimization model is to maximize the expected profit generated by selling recycled materials to the secondary markets and to make the landfilling option less attractive. In comparison with the state-of-the-art stochastic optimization models in this area, which mainly focus on the expected optimal value, this paper emphasizes the role of source separation of recyclable materials to increase the productivity level at the collection centers (CC). Indeed, the quantity of materials collected from the supply sources and the recycling rates at the CC are the primary sources of uncertainty considered in this study. A Sample Average Approximation (SAA) procedure is developed to solve the stochastic model and perform sensitivity analyses on the number of supply sources, the sample size and the level of uncertainty targeting the random parameters. Managerial implications are discussed through a case study from the demolition industry in the province of Quebec, Canada. Although the use of source separation centers (SSC) improves the network performance in both rural and urban zones, the additional flexibility provided by these platforms reaches its best efficiency in the case of high-density urban areas. Finally, the results suggest significant RLND adjustments that lead to an increase in the average profit by 17.6% and recycle around 29% of additional building materials waste from demolition.

Keywords: Network design; Reverse logistics; SAA procedure; Source separation centers; Stochastic optimization.