An Outer Approximation Algorithm for Capacitated Disassembly Scheduling Problem with Parts Commonality and Random Demand

Disassembly scheduling has attained increasing attention in the academic community of reverse logistics. This paper studies a capacitated multi-item multi-period disassembly scheduling problem with parts commonality and random demand. The problem is formulated as a mixed integer nonlinear program (MINLP) with chance constraints. The objective function of the model is to minimize expected total cost, including set-up cost, start-up cost, procurement cost, and expected holding inventory cost. A chance constraint is considered to probabilistically ensure the satisfaction of random demand. Based on the convexity of the proposed model, an outer approximation (OA) algorithm is developed to obtain optimal solutions. Closed-form formulations and numerical experiments are conducted when the demand follows normal distribution. Computational results reflect that the proposed OA algorithm significantly outperforms Bonmin, which is a well-known MINLP solver. Sensitivity analysis reveals practical managerial insights associated with the service level, production capacity, start-up cost, ratio of commonality, and demand deviation. And, a case from a valve maker is presented to demonstrate the application of the research in practice. Finally, conclusions are drawn and future research directions are discussed.