基于相似性自适应学习的遗传算法在无功优化中的应用

针对常规遗传算法(GA)的不足,提出了一种改进的遗传算法—基于相似性自适应学习的遗传算法,为提高遗传算法的计算速度、收敛性和全局最优搜索能力,采取了以下改进措施:①针对遗传算法产生新解无序,提出邻域搜索策略;②为提高算法的搜索效率和效果,按适应值相似性对个体分级、加速;③为提高收敛速度,提出了邻域收缩策略。将改进遗传算法应用于电力系统进行无功优化,在收敛速度和全局收敛性与常规遗传算法进行了比较,结果表明改进遗传算法的有效性。

A genetic algorithm based on similitude frame of evolutionary computation applied in reactive power optimization

Aiming at the shortage of general genetic algorithm,a modified genetic algorithm based on similitude frame of evolutionary computation is proposed. The modified step that is used to improve calculation speed and astringency and entire optimum search ability is given. First,aiming at GA producing no order new individual,the near field search strategy is proposed. Second,according to adaptive value comparability the individual is classified and accelerated for improving search efficiency and effect. Third,the near field shrinkage strategy is proposed for improving rate of convergence. The improved algorithm is used in reactive power optimization of distribution network. Compared with the rate of convergence and the entire astringency of the general and modified genetic algorithm. The results from calculation examples show that the modified method is effective.