用于多目标无功优化的自适应遗传算法

引入了自适应权重和因子及自适应罚函数的概念,提出了一种自适应遗传算法,将其应用于多目标无功优化问题的求解中。该算法能保证寻优方向的多向性,并能避免模糊隶属度算法耗时过长的缺陷。在寻优过程中,自适应罚函数法能有效利用不可行解的有用信息,对不可行解进行适度惩罚。IEEE14节点系统的算例结果表明所提出的算法是解决多目标无功优化问题的有效方法。

A Self-Adaptive Genetic Algorithm for Multi-Objective Reactive Power Optimization

Based on the factor of self-adaptive weight sum and self-adaptive penalty function, a self-adaptive genetic algorithm is proposed and applied to solve multi-objective reactive power optimization. By use of the proposed method the multidirectional property of searching can be ensured and the defect of fuzzy membership algorithm, namely the overlong computing time, can be avoided. During searching process the self-adaptive penalty function can effectively utilize the available information in infeasible solution and appropriately punish the infeasible solution. Results of IEEE 14-bus testing system show that the proposed algorithm is an effective method for multi-objective reactive power optimization.