基于改进遗传算法的配电网网架规划

Distribution Network Planning Based on Improved Genetic Algorithm

针对传统的遗传算法在求解过程中出现收敛速度慢、早熟现象等问题,引入单亲遗传算法;以网架线路年综合费用最小为优化目标建立配电网网架规划的数学模型,运用改进的单亲遗传算法消除了常规遗传算法中对网络可行性破坏严重的双亲交叉算子,同时针对单亲遗传算法在染色体选择、基因操作、收敛准则等方面的不足,采用最优保留策略和两两竞争相结合的染色体选择方法,以及最优个体的最少保留代数为收敛准则,提高了算法的局部搜索效率和全局优化能力,求得模型优化解;仿真算例验证了该方法的可行性和有效性。

Partheno-genetic algorithm （PGA） is introduced to solve the problems of slow and premature convergence, etc. in traditional genetic algorithm （GA） during solution. A mathematical model of distribution network structure planning is established aiming at the minimum annual comprehensive cost of the network. Improved PGA is used to eliminate the parental cross operator in conventional GA that seriously damages the network feasibility. In view of the shortcomings of PGA in such aspects as chromosome selection, genetic manipulation and convergence criterion, the chromosome selecting method combining optimum preservation strategy and pairwise competition is adopted, the least maintaining algebra of the best individual as convergence criterion, thus improving local searching efficiency and global optimizing ability of the algorithm and obtaining the optimal solution of the model. The simulation results verify the feasibility and validity of this method.