Adaptive Genetic Algorithm for Daily Optimal Operation of Cascade Reservoirs and its Improvement Strategy

Kan Yang,Jiao Zheng,Min Yang,Ran Zhou,Guoshuai Liu

DOI: https://doi.org/10.1007/s11269-013-0403-3

IF: 4.426

2013-08-09

Water Resources Management

Abstract:For the specialty of cascade reservoirs optimization and the premature convergence of GA, several improvement strategies are presented in this paper. Firstly, solution space generation method is found application to generate feasible initial population. Secondly, chaos optimization is adopted to optimize initial population. Thirdly, new selective operators, trigonometric selective operators, are proposed to overcome the fitness requirement of non-negative and to maintain the diversity of population. Fourthly, adaptive probabilities of crossing and mutation are adopted in order to improve the convergence speed of GA. Besides, elitist strategy is used to ensure that the best individual can be remained in each generation. Furthermore, the performance of these proposed improvement strategies was checked against the historical improvement strategies by simulating optimal operation of Three Gorges cascade reservoirs premised on historical hourly inflows, and the comparison yields indications of superior performance. In these proposed improvement strategies, trigonometric selective operators are feasible and effective for optimizing operation of cascade reservoirs. These new selective operators could help GA to find a more excellent solution in the same algebra, and the performance of convergence speed is advanced. Adaptive probabilities of crossing and mutation have better performance than other improvement strategies, such as annealing chaotic mutation and simulated annealing of large probability of mutation, because this method realizes the twin goals of maintaining diversity in the population and advancing the convergence speed of GA.

water resources,engineering, civil