改进的多目标遗传算法在配电网规划中的应用

Application of Improved Multi-objective Genetic Algorithm in Distribution Network Planning

针对配电网规划复杂多目标优化问题,建立以经济性和可靠性为目标的配电网规划模型,提出改进的多目标遗传算法,采用分解-协调思想将复杂多变量的规划问题分解成多个子问题,分别对各子问题优化,最终达到全局优化的目的。该算法通过精英选择和个体迁移策略提高了收敛速度,将一种新型编码方式应用于此算法使配电网自然呈辐射状,同时在精华种群中加入裁剪算子以提高搜索效率。与常规算法相比较,该算法收敛性好,最后通过算例表明该算法的有效性。

The distribution network planning is a multi-objective optimization problem, the distribution network planning model considering economy and reliability is built and an improved multi-objective GA is proposed in the paper. The algorithm according to the decomposition-coordination principle is decomposed into a number of sub-problems which are optimized by a single population. The elitist selection and individual migration are used to improve searching speed. The algorithm uses a new code which can make the network radial, and reduced operator in the elitist population is used to search effectively. Simulation shows that compared with GA, the algorithm can avoid the partial optimal solution.