The 0/1 Multi-objective Knapsack Problem Based on Regional Search

A novel evolutionary algorithm is proposed in this paper. The presented algorithm uses regional search strategy to solve MOKP. By this way, the proposed algorithm reduces the computational complexity and accelerates the speed of convergence. This paper uses the greedy repair strategy to handle infeasible individuals during the evolution process. For making the strategy reasonable, we only consider the weight of items in knapsacks which violate the constraint. The experimental results of 0/1 MOKP, with nine testing instances, indicate that the proposed algorithm is highly competitive and can be considered as a viable alternative.