BPSOBDE: A Binary Version of Hybrid Heuristic Algorithm for Multidimensional Knapsack Problems

Abstract

A hybrid heuristic algorithm, named BPSOBDE, is proposed for solving multidimensional knapsack problems (MKPs), in which the basic binary particle swarm optimization (BPSO) is combined with a binary differential evolution (BDE) to maintain the diversity of the swarm and makes it more explorative, effective and efficient. BPSOBDE is tested through computational experiments over suites of benchmark problems and the obtained results are compared with those of BPSO and some modified versions of BPSO. The experimental results show that BPSOBDE can successfully locate the exact solutions of all test problems. The comparison results indicate that BPSOBDE is a competitive heuristic algorithm.