Reduction of the Criteria System for Identifying Effective Reservoirs in the Joint Operation of a Flood Control System

The accurate and efficient identification of effective reservoirs plays an important role in the real-time flood control operation of multireservoir systems. The redundancy of the initial criteria system affects the efficiency of identifying effective reservoirs. In this paper, a criteria reduction model based on the equivalence relation of rough set theory is proposed to solve this problem. Correspondingly, the decision rules are derived for identifying effective reservoirs. In addition, the reduction result of rough set theory is compared with the results of correlation analysis-principal component analysis (CA-PCA), vague set theory and back-propagation (BP) neural network methods. The proposed methodology is applied to the flood control system in the upper and middle reaches of the Huaihe River basin in China. The results show that the dimensionality of the initial criteria system is reduced and that the reduction result is reasonable. The proposed reduction model is an effective tool to reduce the initial criteria system for effective reservoir identification in real-time flood control operation.