基于环境Pareto支配选择差分进化算法的舰船电网重构

为提高差分进化算法在舰船电网重构中寻找最优解的准确性，避免陷入局部最优，采用一种环境Pareto支配的选择策略，对变异后得出的可行解与优秀不可行解同时进行选择，根据改进的选择策略对优秀不可行解加以利用.针对舰船电网重构的离散多目标优化问题，采用0,1,2离散编码和无参数变异算子进行离散操作，并利用混沌初始化提高种群多样性.仿真实验表明，相比其他算法，本文算法具有更好的故障恢复方案、收敛性和稳定性，并能有效避免陷入局部最优.
In order to improve the accuracy of the optimal solution and avoid local optimization in shipboard power grid reconstruction, an improved differential evolution algorithm was proposed by using the environment Pareto dominated selection strategy, the feasible solution and the excellent infeasible solution after mutation were selected, and then the excellent infeasible solution was utilized based on improved selection strategy. Aiming at the discrete multiobjective optimization problem of shipboard power grid reconstruction, the discrete operation was conducted by using the 0, 1, 2 discrete codes and nonparametric mutation operator, and the chaotic initialization was adopted to improve the population diversity. Results show that compared with other algorithms, the improved differential evolution algorithm has better failure recovery scheme, convergence and stability and effectively, as well as avoiding local optimization.