Study On Reservoir Optimal Operation Based On Genetic Algorithm

Reservoir operation is in order to realize the water resource redistribution, and achieves the coordinated development among the industry, agricultural, human life and environment, which depend on advanced technology and optimization plan. According to this aim, this dissertation takes Baojixia etc five reservoirs as research objects, and has studied genetic algorithm-based reservoir operation methods with combination of optimization technology and database technology, in which single reservoir and reservoir group are stressed. Finally, Weihe River reservoir operation system is developed, and which provide the technical support for modern management of reservoir operation in future.The main contents of this research are as follows:(1) On the basis of Weihe River distribution, node type, logic relationship and interactive way of fluent among water source is studied, and valley topological structure is abstracted. According to node connection relationship, valley conception model is established by use of graph structural style. In these foundations, the recursion theory of depth first traversal, database technology and visual programming technology are employed to computer memory, which is expressed to node relations.(2) According to reservoir operational mechanism, water resource assignment and reused principle, Valley reservoir operation mathematical models are established, which unify the reservoir operation characteristic and take the system engineering as theory.(3) On the basis of standard genetic algorithm existent question, reservoir optimization method is prompted which based on adaptive genetic algorithm. Algorithm solution model of Single reservoir and reservoir group are separately designed and realized by reservoir storage capacity discretization.(4) In order to validate the algorithm superiority, which used in the dissertation, Dynamic Programming (DP), Standard Genetic algorithm (SGA) and Adaptive Genetic algorithm (AGA) are applied to reservoir optimization operation. Experimental results prove that Genetic Algorithm could guarantee to find the global optimum quickly, to save a lot of computer memory, which could avoid the dimension problem, that SGA could adjust the parameters adaptively according to the individual fitness and dispersion degree of population, which solves the convergence and prematurity problems well occurred in SGA.