Landscape analysis and scatter search metaheuristic for solving the uncapacitated single allocation hub location problem

Abstract

In this paper, solution space landscapes of the uncapacitated single allocation hub location problem (USAHLP) for the well-known CAB and AP benchmark datasets are investigated and analysed through several statistical criteria. The analyses show that both datasets have rugged landscapes, and the optimal solutions are concentrated in the search space of the CAB dataset while they are uniformly distributed in the search space of the AP dataset. It follows that in order for a method to find good solutions to the USAHLP, it should perform both proper exploitation and exploration of the workspace. Based on the landscape analysis, an efficient scatter search-based heuristic method called SSUHLP is tailored for solving the USAHLP. The performance of the SSUHLP is evaluated by solving all problems of the CAB and AP datasets and comparing them with some existing algorithms in the literature.