求解舰船电力系统网络重构的贪婪DPSO算法

t  目的   t为了解决现有约束多目标优化算法在求解船舶微电网重构时收敛性和分布性不佳的问题，提出一种基于两阶段差分进化（TSDE）算法的约束多目标优化方法。t  方法   t第1阶段采用双种群混合法（即自适应罚函数法和可行性法则）来处理约束条件；第2阶段将第1阶段产生的双种群合并为单种群，再采用可行性法则解决约束优化问题；最后，在不同的阶段采用不同的精英选择策略和改进无参数变异算子，从而进一步优化差分进化算法。t  结果   t根据算例仿真结果：在故障1和故障2工况下，TSDE算法求得的最小负荷失电量分别比基于混沌迁移及无参数变异差分进化（CMPMDE）算法和基于环境Pareto支配选择差分进化（EPDSDE）算法降低了185 A和940 A；在故障1工况下，TSDE算法的最少开关操作数比CMPMDE算法多1次，与EPDSDE算法相同；在故障2工况下，TSDE算法的最少开关操作数比CMPMDE算法和EPDSDE算法均少1次。t  结论   tTSDE算法求得的最优非劣解集更接近真实的Pareto前沿且分布较为均匀，在满足重构时间要求的前提下，该算法可以更好地保证船舶的安全稳定运行。

Greed DPSO algorithm for network reconfiguration of shipboard power system

t  Objectives   tIn order to solve the problem of poor convergence and distribution of the existing constrained multiobjective optimization algorithms in solving the ship micro-grid reconfiguration, a constrained multiobjective optimization method based on two-stage differential evolution (TSDE) algorithm is proposed.t  Methods   tFirstly, in the first stage, the two-population hybrid method(i.e. self-adaptive penalty function method and feasibility rule)was used to deal with the constraints. Secondly, in the second stage, the two populations generated in the first stage were merged into a single population, and the feasibility rule was adopted to solve the constrained optimization problem. Finally, different elitist selection strategies and improved non-parametric mutation operators were adopted in different stages to further optimize the differential evolution algorithm.t  Results   tThe simulation results show that the minimum load loss obtained by TSDE algorithm under the fault 1 and the fault 2 is 185 and 940 A lower than that of chaotic migration and parameterless mutation differential evolution (CMPMDE) and environment pareto dominated selection differential evolution(EPDSDE), respectively. The minimum switching operands obtained by the TSDE algorithm are 1 time more than that of CMPMDE algorithm under the fault 1, and are the same as that of EPDSDE algorithm. Under the fault 2, the minimum switching operands of the proposed algorithm are 1 time less than those of CMPMDE algorithm and EPDSDE algorithm.t  Conclusions   tThe set of optimal non-inferior solutions obtained by TSDE algorithm is closer to the real Pareto frontier and distributes more evenly, so the method can ensure that the ship is operated safely and steadily when the reconfiguration time is satisfied.