采用退火遗传算法的配电网优化方法  

Optimization of power distribution with annealing genetic algorithm

作　　者：张宁 马孝义 杨震宇 黄文文 

机构地区：[1]西北农林科技大学水利与建筑工程学院电气工程系,陕西杨凌712100

出　　处：《重庆大学学报（自然科学版）》2014年第2期75-80,97,共7页Journal of Chongqing University

基　　金：国家自然科学基金项目资助(50879072);国家级农业水工程实验教学示范中心建设项目资助(AWEDF305);西北农林科技大学大学生创新创业训练计划项目(2201310712077)

摘　　要：在分析多约束多目标配电网优化模型基础上,针对传统遗传算法求解时静态稳定性和鲁棒性不强、爬坡能力弱、收敛速度慢等问题,引入网络拓扑简化方法,决策变量多参数编码和退火动态罚因子处理约束,提出了基于退火遗传算法的配电网模型求解方法。通过实际农网决策工程验证表明,退火遗传算法适应度爬坡能力明显加强,稳定于1的几率提高,收敛于全局最优次数增多,运行时间缩减,而且综合费用比传统遗传优化方案减少了近50万元,从而达到了进一步优化的目的。The traditional genetic algorithm easily converges to slow speed,leads to a poor climbing ability and shows weak static stability and robustness when seeking a solution.On the basis of analyzing multi-constrained and multi-objective optimization model of power distribution network,this paper describes a method of adopting the simplified network topology method,the multi-parameter decision variable coding and the dynamic penalty functions for model constraints and the annealing genetic algorithm.When the method is used for the actual decision project in rural power distribution network,the results show that the proposed method can obviously strengthen climbing ability,improve the stability chance at 1 ,increase the times of converging to global optimization and cut running time.Meanwhile,the comprehensive cost is nearly 0.5 million yuan cheaper than traditional genetic optimization scheme,which,therefore,achieves further optimization.

关 键 词：配电网 规划 经济性 算法 退火 

分 类 号：TM727.1[电气工程—电力系统及自动化]