A multi-objective risk management model for real-time flood control optimal operation of a parallel reservoir system

Many uncertainties are associated with real-time reservoir flood control operation and introduce risks for decision making. A multi-objective risk management model is proposed to optimize the real-time flood control operation of a parallel reservoir system during flood season. The model considers the combined impact of reservoir inflow and lateral inflow uncertainties and flood control operation risks. It consists of three key components: a risk optimal operation submodel, a risk assessment submodel, and an optimization submodel. The risk optimal operation submodel takes into account the uncertainties and establishes an operation model that considers two competing objectives of minimizing the risk of upstream flooding and minimizing the risk of downstream flooding. The risk assessment submodel calculates the risks based on a stochastic differential equation (SDE). The final optimization submodel embeds the risk assessment model in the risk optimal operation model and is solved by the Non-dominated Sorting Genetic Algorithm-III (NSGA-III). The proposed methodology is applied to a real flood control system in the middle reaches of the Huaihe River basin in China. The results show that the developed multi-objective risk management model can provide operation schedules that satisfy flood control objectives and simultaneously minimize the overall risks. The Pareto front of the proposed objectives demonstrates the tradeoff among the competing objectives. The model can be used as a decision-making tool for conducting risk management for real-time reservoir flood control operation during flood season. The decision makers can choose the operation schedule according to their risk preference and updated hydrological information.