Adaptive variable neighborhood search for solving multi-objective facility layout problems with unequal area facilities

In this paper, we report the results of our investigation of an evolutionary approach for solving the unequal area multi-objective facility layout problem (FLP) using the variable neighborhood search (VNS) with an adaptive scheme that presents the final layouts as a set of Pareto-optimal solutions. The unequal area FLP comprises a class of extremely difficult and widely applicable optimization problems arising in diverse areas and meeting the requirements for real-world applications. The VNS is an explorative local search method whose basic idea is systematic change of neighborhood within a local search. Traditionally, local search is applied to the solutions of each generation of an evolutionary algorithm, and has often been criticized for wasting computation time. To address these issues, the proposed approach is composed of the VNS with a modified 11-opt local search, an extended adaptive local search scheme for optimizing multiple objectives, and the multi-objective genetic algorithm (GA). Unlike conventional local search, the proposed adaptive local search scheme automatically determines whether the VNS is used in a GA loop or not. We investigate the performance of the proposed approach in comparison to multi-objective GA-based approaches without local search and augmented with traditional local search. The computational results indicate that the proposed approach with adaptive VNS is more efficient in most of the performance measures and can find near-optimal layouts by optimizing multiple criteria simultaneously.