Abstract:

Aiming at the multiple fuzzy characteristics of vehicle routing problem with hard time windows in battlefield resources distribution, a multi-objective fuzzy expected model is designed based on the fuzzy credibility theory, and an improved constrained multi-objective particle swarm optimization is proposed. The algorithm puts forward a particle encoding technique according to phase-space, and designs a non-dominated solutions constructing method which can store excellent infeasible solutions. In order to improve the convergence performance, a hybrid diversity strategy based on adaptive grids and crowding distance is used to maintenance the non-dominated solutions, and the individual guide updating mechanism is improved. Meanwhile, local search and mutation operators are introduced to avoid the algorithm premature convergence. The rationality of the proposed model and effectiveness of the algorithm are verified by computational experiments.