Simultaneous Optimization of Operating Rules and Rule Curves for Multireservoir Systems Using a Self-Adaptive Simulation-GA Model

Abstract

This study introduces a simulation-optimization model for deriving an operating policy for multireservoir systems. Two adjustable monthly rule curves are introduced to each reservoir in the system. The applied rule curves divide the reservoir volume into three zones in each within-year period. For each zone, a release coefficient is specified to indicate releases from the reservoir as a function of the available storage and time of the year. To obtain optimum rule curves and release rules, a self-adaptive genetic algorithm (GA) is developed to maximize the system’s hydropower production, subject to the system’s physical constraints as well as a desirable reliability to satisfy water demands. To evaluate the objective function and constraints, a simulation model based on the network flow technique and linear programming is developed and coupled to the GA. The model is applied to a real three-reservoir system in Iran and the results are discussed and compared to the standard operation policy (SOP).