基于遗传算法的机组组合研究

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／变异率的遗传算法和机组组合计算程序，并对10机系统和110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，为遗传算法应用到机组组合问题的实用化研究奠定了坚实的基础。

STUDY OF GA-BASED UNIT COMMITMENT

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. A good foundation is settled for further practicable study of GA based UC.