Design and analysis of collection systems for incentive-dependent product recovery

Abstract

In this thesis, we propose three reverse logistics models in which we address the problem of locating collection centers of a profit seeking company that aims to collect used products (cores) from product holders via a pick-up strategy. Firstly, we formulate a mixed-integer nonlinear facility location-allocation model to find both the optimal locations of collection centers and the optimal quality dependent incentive values to be paid to product holders for returning their cores. Furthermore, we elaborate on two bilevel programming formulations to model the relationship between the government and the company engaged in core collection operations. Since the company seeks only economic profitability, the collected amounts may not be aligned with the target collection rate imposed by the government. In both models, the government pays a unit subsidy to the company for each core collected. The two models differ from each other by the attitude of the government towards the company as being supportive or legislative. We propose heuristic methods to solve medium and large size instances. For the company’s problem, the main loop of the method is based on tabu search performed in the space of collection center locations and Nelder-Mead simplex search is called to determine the best incentives and the corresponding net profit. For the government’s problem, we propose a solution approach based on Brent’s method, which is a root finding method. Our heuristics obtain good results in all models compared to the results of commercial solvers.