Accelerating genetic algorithm based on optimization of feasible search space and its application in cascade reservoir optimal operation analysis

The cascade waterpower development is an important component of the world energy development, as an typical non-engineering measure of cascade waterpower development, combined optimal operation of cascade hydropower stations, involving widely discipline categories and department interests, is essentially an decision-making problem of large complex system, which has a profound theoretical and practical guiding significance for realizing the coordinated and sustainable development of regional water resources, social economy and ecological environment. As the researching difficulties of large complex constraints and curse of dimensionality when applying the swarm intelligent optimization algorithm in optimal operation of cascade hydropower stations, the accelerating genetic algorithm (AGA), in which the optimizing space of the variables are shrinking gradually with the increase of the iteration times, was introduced in the research of optimal operation of cascade hydropower stations, and the feasible decision space for the optimization variables of different station and period were formed based on the principles of classification and reverse iteration, and then the accelerating genetic algorithm based on feasible search space optimization and its application in the cascade waterpower development were discussed in this paper. Finally, applying the method in the combined optimal operation example of Wujiang cascade hydropower stations, which has a good representative of hydraulic and power characteristics in China, and the result shows that, the accelerating genetic algorithm based on feasible search space optimization has strong self-adaptability and global optimization ability for solving the complex constraints and improving the optimization efficiency, and the average annual energy output is 289.67×10
⁸
kW•h, with the increasing extent of 2.60% when comparing with the designing values of the Wujiang cascade stations. So, it is obvious that applying the accelerating genetic algorithm based on feasible search space optimization in the optimal operation research of cascade hydropower stations is feasible and reasonable, and the research results would provide scientific and effective decision basis for the centralized dispatching and operation of cascade hydropower stations, and also enrich and develop the theoretical and practical research system of regional waterpower development and utilization.