A memetic algorithm for mixed-model two-sided disassembly line balancing problem

Abstract

Disassembly lines have emerged to quickly disassembly large amounts of End-of-Life products. In the literature, there are still deficiencies in disassembly line balancing works for large-sized products (e.g. trucks, refrigerators) where plugins such as complete disassembly, mixed-model, and various solution methods will be used for various objectives. The aim of this study is to fill this gap in the literature. In this study, we developed a Mixed Integer Linear Programming (MILP) model that guarantees the optimum result for the Mixed-Model Two-Sided Disassembly Line Balancing (MTDLB) problem, which allows disassembling of different products on the same line and working synchronously on both sides of the line. AND/OR Graph, which provides access to all disassembly precedence diagrams, is used. By selecting only one of the task with OR Successor relation of each task, finding the most suitable disassembly sequence and balancing the disassembly line by assigning the selected tasks to the appropriate stations were handled simultaneously. A Memetic Algorithm (MA) has been designed to achieve near optimum results of large-scale cases. The solution performances of the proposed methods have been tested on a series of test problems against Genetic Algorithm (GA) and Gurobi solver. The computational results show that the MILP mathematical model can be used for the solution of small-scale cases, and MA can be used compared to GA for medium-scale and large-scale cases.