Set pair analysis and BP neural network coupling model for optimal selection of flood control operation

Reservoir flood control operation, which plays an important role in reducing the flood loss, ensuring the safety of life and property and realizing the safe utilization of flood resources, is a complex decision making problem involving flood control, water supply and power generation. For the precise optimal solution of reservoir flood control operation is difficult to be found, thus the researching emphasis is how to identify the optimal scheme from the feasible scheme set. Basing on the establishment of evaluating scheme and index system for reservoir flood control operation, the SPA method can be used to determine the connection number component corresponding to different indexes between the feasible and ideal schemes, and through the weight values of indexes determined by AGA-FAHP, the connection number component would be synthesized into final connection number. Thus the ANN-AGA would be used to evaluate the practical schemes, and by the sorting for the schemes according to their final connection numbers, then the set pair analysis and BP neural network coupling model for optimal selection of flood control operation (named BP-SPA for short) is proposed. The applying results indicate that this method can reflect the differences between the practical and ideal optimal schemes comprehensively, and the evaluating result is reasonable and reliable, so it can be applied to other optimal selection process of complex systems.