An approach to design reverse logistics networks for product recovery

Shortened product lifecycle has resulted in increasing number of obsolete products, leading to growing environmental concerns. Product recovery, which is a way of reducing waste, is gaining popularity in recent years. To effectively implement product recovery activities, a reverse logistics network design that can facilitate reverse flow of used products in an efficient way is required. In this paper, an approach to design reverse logistics network for product recovery, incorporating multi-objective functions, multi-period planning horizons, and uncertainties is presented. This approach consists of both mathematical and simulation models; the former utilizes mixed integer programming (MIP) to model the multi-objective, multi-period problem of network design, whereas the latter is employed to incorporate uncertainties. Due to its complexities, spanning-tree based genetic algorithms are utilized to find non-dominated solutions. The preferred non-dominated solutions are simulated under several scenarios of uncertainties to determine the best-preferred reverse logistics network design.