基于混沌迁移及无参数变异差分进化算法的舰船电力系统网络重构  

Network reconfiguration of ship power system based on chaotic migration and parameterless mutation differential evolution algorithm

作　　者：马理胜[1] 张均东[1] 任光[1] 王俊 

机构地区：[1]大连海事大学轮机工程学院,辽宁大连116026 [2]厦门船舶重工股份有限公司,福建厦门361026

出　　处：《上海海事大学学报》2015年第3期76-81,共6页Journal of Shanghai Maritime University

基　　金：国家自然科学基金(51179102)

摘　　要：为更好地利用差分进化算法对舰船电力系统网络进行重构,提出一种基于混沌迁移及无参数变异的差分进化算法.针对差分进化算法寻找最优解容易陷入早熟的问题,引入一种基于混沌迁移的并行进化策略.该策略将原有种群分为多个子种群,进行并行进化.在优化过程中引入混沌迁移序列引导个体迁移,利用混沌的遍历性和随机性,保证子种群间能高效地进行信息交换.针对电力系统网络重构中的0,1,2编码方式在解码中信息丢失问题,提出一种无参数变异算子.这个算子能使算法结构简单、利于运算.最后利用混沌序列初始化种群和Pareto选择策略提高舰船重构效率.仿真实验表明,改进的算法具有更好的故障恢复方案,能有效避免差分进化算法在求解电力系统网络重构时的早熟问题.To better use Differential Evolution （DE）algorithm to reconfigure the ship power system net-work,a DE algorithm based on chaotic migration and parameterless mutation is proposed. To address the problem of premature resulting from DE algorithm while searching for the optimal solution,a parallel evo-lution strategy based on chaotic migration is proposed. The strategy divides the original population into several sub-populations,and then carries on parallel evolution. In the optimization process,the chaotic migration sequence is introduced to guide individual migration,where the ergodicity and randomness of chaos make sure the efficient information exchange among the sub-populations. To address the information lost in the decoding process where the 0,1,2 encoding mode is used to reconfigure the network,a pa-rameterless mutation operator is proposed. The operator makes the algorithm structure simple and easy to calculate. To enhance the efficiency of reconfiguration,the chaotic sequence is adopted to initialize the population and the Pareto selection strategy is adopted. Simulation experiment results show that the im-proved algorithm can provide a better service restoration plan and can solve the problem of premature in ship power system network reconfiguration.

关 键 词：舰船电力系统 差分进化算法 混沌迁移 无参数变异 网络重构 

分 类 号：U665.12[交通运输工程—船舶及航道工程]