Multi-objective Evolutionary Algorithm Based On Decision Preference For Reservoir Flood Control Operation

As an important part of water conservancy project, reservoir plays an important role in flood control, irrigation, water supply, environment and so on. It has obvious social and economic benefits. At the same time, it is related to the life and property safety and economic sustainable development of the upper reaches of the reservoir. Reservoir is a great event which involves public security, social stability and the development.Reservoir flood control operation implement the flood relief scheme based on the upstream flood inflow situation, using of the reservoir engineering, according to the established schedule. The optimal operation of reservoir flood control includes some constraints, which should take the upstream and downstream flood control objectives into account and also to consider the safety of dams itself. The reservoir flood control operation is a large multi-purpose, multiple constraints, multi-phase, nonlinear optimization problem. In addition, considering the long-term goal namely the irrigation requirement after the flood, the final scheduling water level of the reservoir flood control scheduling should within a certain specified range. The dispatching schemes which can also meet the demand of flood irrigation after flood after guaranteed the reservoir upstream and downstream as well as dam own safe can satisfy is the preference of decision makers. Therefore, the preference model of multi-objective reservoir flood control operation(RFCO) emerges as the times require. The main work of this thesis is as follows:For the optimization problem of reservoir flood control with two objectives including the upstream and downstream safety goal and the dam safety goal, this thesis designs a preference model through weight adjustment method based on the preference information of irrigation demand after flood, and introduced it into the framework of the decomposition based multi-objective evolutionary algorithm(MOEA/D) to form the newly developed MOEA/D-PWA for solving multi-objective RFCO problem. Experimental results on four typical floods at the Ankang reservoir have illustrated that the suggested MOEA/D-PWA can successfully converge to the preferred region. The dispatching schemes obtained by MOEA/D-PWA can significantly reduce the flood peak and guarantee the dam safety as well. The proposed MOEA/D-PWA is an efficient use of computing efforts.For the optimization problem of reservoir flood control with three objectives, which includes the safety goal of the upstream, the downstream safety and the reservoir and the dam itself, and the specific preference model is designed for the three objective optimization through the method of weight adjustment. Similarly, we carried on the simulation experiment to the four typical flood of Ankang reservoir. Experimental results verify the proposed three goals MOEAD-PWA can effective use of computing resources, can significantly reduce the peak ensure dam safety and the safety of the downstream, successfully the aggregation results in the preferred region at the same time which means that irrigation demand after reservoir flood can be satisfied, and compared with the amount of generation by the reservoir flood scheduling scheme from the two-objective optimization MOEA/D-PWA algorithm, we can find that three-objective optimization algorithm can obtain scheduling schemes which can generate greater power generation.