An affinity propagation-based multiobjective evolutionary algorithm for selecting optimal aiming points of missiles

When missiles attack the group targets, the selection of their optimal aiming points is a nonlinear, multi-dimensional and multimodal multiobjective optimization problem. To effectively address this problem, an affinity propagation-based multiobjective evolutionary algorithm called APMO is proposed in this article by introducing an affinity propagation and reproduction utility-based adaptive mating selection strategy named as AMS. In AMS, at each generation, an affinity propagation approach is firstly utilized to discover the neighborhood relationship of solutions. Afterward, parent selections for recombination are conducted on the neighborhoods or the whole population based on a mating restriction probability. Moreover, the mating restriction probability is updated at each generation according to the reproduction utility of the neighborhoods and the whole population over the last certain generations. Comprehensive experiments on benchmark instances denote that the proposed APMO significantly outperforms five popular multiobjective evolutionary algorithms, MOEA/D-DE, RM-MEDA, NSGA-II, SPEA2 and MOEA/D-STM. Practical application proves that APMO is promising to select the better optimal aiming points for missiles.