有限制自适应遗传算法在三峡水电站厂内经济运行中的应用

针对三峡水电站厂内优化运行问题，提出一种改善算法收敛性的有限制自适应遗传算法．为了保证适应度函数的非负性，采用包含解空间初始种群生成和有限制摄动变异的有限制策略．运用个体优劣和群体分散程度自适应调整的交叉和变异概率，保持种群多样性和提高收敛速度．三峡水电站厂内经济运行不同量级的26台机组负荷分配结果表明：有限制策略解决了机组空蚀振动问题且保证了适应度函数的非负性，改善了算法的收敛性；同时，根据个体优劣和群体分散程度进行参数自适应调整的策略对于求解水电站厂内经济运行来说，是一种有效可行改进方式．

Three Gorges Hydropower Station Inner-Plant Economical Operation Based on Limited Adaptive Genetic Algorithm

To deal with Three Gorges hydropower station inner-plant economical operation problem, an improved adaptive genetic algorithm （GA） was proposed to enhance convergence performance of algorithm. Limited strategy, with the feasible solution generation method for generating initial population and the limited perturbation mutation operator, was presented to keep nonnegative of fitness function. And the adaptive probabilities of crossover and mu- tation, according to the individual fitness and group dispersed degree, were introduced to keep diversity in the popu- lation and to advance the convergence speed of GA. The results of Three Gorges hydropower station inner-plant eco- nomical operation with 26 hydro units based on several representative loads demonstrate the effectiveness of the pro- posed approach. Firstly, limited strategy solves cavitation-vibration problem and improves convergence speed by keeping nonnegative of fitness function. Secondly, the adaptive parameter method gives an important reference to optimizing hydropower station inner-plant economical ooeration.