摘要
以基于实测的电力系统综合负荷建模为应用对象,探讨遗传算法的运行机理,分析遗传算子的不同搜索能力。指出决定遗传算法性能的关键因素是种群多样性,得出了种群多样性与算法参数的关联约束。从理论分析上给出遗传参数的设定规则,深入研究遗传算法中种群规模、交叉、变异概率及其控制策略,以及初始种群参数区间等遗传算法关键操作参数对算法性能的影响规律,给出合理的种群规模和参数初始区间,提出与群体进化程度指标相关的自适应调整交叉概率和变异概率策略。研究结果表明,合理的参数组合是挖掘遗传算法潜能的关键,可提高遗传算法运行效率、克服早熟及尽量减小模型参数分散性。
Combining with the application of GA（Genetic Algorithm） in power system load modeling based on measured field data,the operation mechanism of genetic algorithm is discussed and the search ability of different genetic operators is analyzed. The population variety is the key factor determining GA performance and the relative restraints between population variety and algorithm parameter are obtained. Rules of genetic parameter setting are provided from theoretical analysis,and the population scale,crossover probability,mutation probability and their control strategies are studied ,as well as the influence of key operation parameters ,such as initial parameter ranges ,on the performance of GA. Reasonable population scale,initial parameter ranges are put forward,and the self-adaptive modulation strategies for crossover and mutation probabilities,which are relative to innovation degree,are proposed. The result of study indicates that appropriate parameter combination is the key to excavate the potential of GA,which improves its operational efficiency,avoids precocity and reduces the dispersiveness of model parameters.
出处
《电力自动化设备》
EI
CSCD
北大核心
2006年第5期23-27,共5页
Electric Power Automation Equipment
基金
高等学校骨干教师资助计划项目(教技司[2002]65号)
湖南省教育厅重点项目(湘教通[2001]197号)
关键词
电力系统
负荷建模
遗传算法
控制参数
参数区间
electric power system
load modeling
genetic algorithm
control parameter
parameter range
