基于遗传算法的机组组合研究  

STUDY OF GA-BASED UNIT COMMITMENT

作　　者：汪峰[1] 朱艺颖[2] 白晓民[2] 

机构地区：[1]国家电力公司发输电运营部,北京市100031 [2]中国电力科学研究院,北京市100085

出　　处：《电力系统自动化》2003年第6期36-41,共6页Automation of Electric Power Systems

摘　　要：针对遗传算法应用于机组组合问题的具体实现技术进行了深入的研究 ,实现了采用不同采样空间、不同选择策略、不同适值函数和不同交叉率 /变异率的遗传算法的机组组合计算程序 ,并对1 0机系统和 1 1 0机系统的仿真计算进行了分析和比较。结果表明各种不同实现技术的遗传算法应用到机组组合问题具有不同程度的有效性 ,另外 ,遗传算法的不同实现技术对收敛时间、收敛代数和收敛值也有较大的影响。文中在计算研究的基础上提出了适用于机组组合问题的遗传算法的具体实现技术 。Unit commitment (UC) based on the genetic algorithm (GA) with different implementation techniques, such as different sampling space, different selection schemes, different fitness scaling methods and different crossover/mutation rate, is developed and implemented in 10-unit system and 110-unit system. This paper thoroughly studies the effects on convergence time, convergence generation and convergence value of GA due to different implementation techniques. It indicates that the algorithm is effective in a certain degree, and reveals that different implementation techniques have great effect on convergence time, convergence generation, and convergence value of GA-based UC.

关 键 词：发电计划 机组组合 遗传算法 

分 类 号：TM769[电气工程—电力系统及自动化]