CAJ | 학술논문

针对水库群发电优化调度的高维性和复杂非线性,提出了一种收敛性全面改善的改进自适应遗传算法.改进算法采用了整体改进策略:初始种群生成方面,运用加入混沌优化的初始种群解空间生成法,改善初始种群质量;选择运算方面,采用对适应度函数进行非线性转换的三角函数选择算子,保持种群多样性;交叉和变异运算方面,采用随个体优劣和种群分散程度自适应调整的交叉概率和变异概率,提高算法收敛性.以典型入库流量下三峡梯级水库发电优化调度为实例,对比了上述整体改进策略和传统改进策略,结果表明:整体改进策略和传统改进策略相比,在克服遗传算法早熟和提高算法收敛性能方面有一定的优势,改进自适应遗传算法适用于求解水库群发电优化调度问题.
침대수고군발전우화조도적고유성화복잡비선성,제출료일충수렴성전면개선적개진자괄응유전산법.개진산법채용료정체개진책략:초시충군생성방면,운용가입혼돈우화적초시충군해공간생성법,개선초시충군질량;선택운산방면,채용대괄응도함수진행비선성전환적삼각함수선택산자,보지충군다양성;교차화변이운산방면,채용수개체우렬화충군분산정도자괄응조정적교차개솔화변이개솔,제고산법수렴성.이전형입고류량하삼협제급수고발전우화조도위실례,대비료상술정체개진책략화전통개진책략,결과표명:정체개진책략화전통개진책략상비,재극복유전산법조숙화제고산법수렴성능방면유일정적우세,개진자괄응유전산법괄용우구해수고군발전우화조도문제.
For the high dimension and complex nonlinear characteristic of multi-reservoir system generation optimization, a comprehensive convergence improved adaptive genetic algorithm (AGA) is proposed. In this algorithm, initial population solution space generation method with chaos optimization is used to improve the quality firstly. Secondly, trigonometric selective operator, which does nonlinear transformation on fit?ness,is adopted to maintain the diversity of population. Finally,adaptive parameters of crossover and muta?tion, depending on the fitness value of individual and the dispersion degree of population, are adopted to improve the convergence speed of GA. Furthermore, the performance of these proposed improvement strate?gies are checked against with the historical improvement strategies by optimal operation of the Three Gorges cascade reservoirs premised on historical hourly inflows. The results show that the proposed improvement strategies have advantages in overcoming GA premature and improving convergence performance, and this improved AGA is feasible and effective for optimal generation dispatching of multi-reservoir system.