基于改进小生境遗传算法的配电网规划

Distribution Network Planning Based on Improved Niche Genetic Algorithm

提出了改进小生境遗传算法并将其应用于配电网规划,建立了综合考虑网损、变电站馈线建设费用和网络年维护折旧费的配电网规划数学模型,算法主要做了以下改进:设置个体之间的距离判别标准为动态函数,保持了群体的多样性;采用最优保存策略保证算法的收敛性;采用局部最优徘徊策略作为辅助终止原则,避免陷入局部最优的情况。在此基础上,基于图论知识和搜索技术给出了包括孤链、孤岛、闭环的修复方案,将遗传操作产生的非辐射性网络修复成辐射性网络。算例结果表明了该算法的实用性和有效性。

This paper presented an improved niche genetic algorithm, and applied it in the distribution network planning. A distribution network model was established, which took into consideration the loss、feeder substation construction cost and network annual maintenance depreciation. The improvements containedinthealgorithmwasasfollows：thediscriminatingstandard length for distances amongindividuals is set as the dynamic function tokeepthediversityofcolonies;theelitiststrategyisadoptedtoensure theconvergenceofthealgorithm;thecirclingstrategyoflocalextreme values is used as auxiliary termination principle to avoid falling into local optima. On this basis, a restoration scheme including the isolated chain, isolated nodes and closed loop based on the graph theory and search technology, restores the nonradial network generated by genetic operation to radial network. The calculation examplesshowthatthismethodispracticalandeffective.