Genetic Algorithms for Optimal Reservoir Dispatching

The fundamental guidelines for genetic algorithm to optimal reservoir dispatching have been introduced. It is concluded that with three basic generators selection, crossover and mutation genetic algorithm could search the optimum solution or near-optimal solution to a complex water resources problem. Alternative formulation schemes of a GA are considered. The real-value coding is proved significantly faster than binary coding, and can produce better results. Sensitivity of crossover probability and mutation probability are also analyzed in this paper. Results from genetic algorithm with real-value coding are compared with those from other optimal methods. The results demonstrate that a genetic algorithm can be satisfactorily used in optimal reservoir problems, and it has potential in application to complex river systems. Copyright Springer Science + Business Media, Inc. 2005