改进的混合遗传算法在配电网规划中的应用

Application of Improvement Hybrid Genetic Algorithm to Distribution Network Planning

利用遗传算法解决NP问题(非确定性多项式问题)的良好能力、模拟退火算法在当前点邻域内搜索最优解的能力和禁忌搜索算法较快的搜索速度,提出了改进的混合遗传算法,即将模拟退火选择算子和禁忌搜索变异算子应用到遗传操作中,提高了种群选择的有效性和遗传算法局部搜索能力,避免了单一遗传算法中收敛速度慢和早熟现象的产生。并将改进的混合遗传算法应用到河南省北部A地区电网规划中,对水平年电力网架进行了优化,结果满足电力系统经济性和可靠性的要求,运行实践证明该算法效果较好。

Based on the good abilities of genetic algorithm （GA） for solving NP problems, searching optimal solution in current point＇s neighborhood of simulated annealing （SA） and faster convergence of Tabu Search （TS）, this paper pro poses an modified hybrid genetic algorithm. Select operator in SA and mutation operator in TS are applied to GA, which improves the effectiveness of selecting population and the local search abilities in GA, and avoids slow convergence and premature phenomena in GA. And then an improved hybrid genetic algorithm is applied to regional distribution network planning in North area of Henan Province. The power network frame in level years is optimized. The optimization results meet the requirements of power system economy and reliability. The effectivess of the proposed method is proved by the operation practice.