Two-Stage Metaheuristic Mixed Integer Nonlinear Programming Approach to Extract Optimum Hedging Rules for Multireservoir Systems

Abstract

Hedging policy is an applicable strategy for water resources systems with storage reservoirs to reduce the damage resulting from extended droughts. Optimizing the operation of parallel and cascade multireservoir systems is a challenging task because of its many complexities. This study proposes a two-stage approach for reaching both operating rule curves and rationing coefficients for multireservoir systems. In the first stage, initial solutions of the main problem are achieved by coupling the mixed-integer nonlinear programming (MINLP) with the particle swarm optimization algorithm. The achieved solutions are then adjusted in the second stage by implementing a distributed simulation–optimization model that efficiently models the system features. The proposed approach has been applied to gain optimal rule curves and rationing coefficients for the reservoirs of the Great Karun multireservoir system in southwestern Iran. The proposed model outperformed the other two—i.e., the distributed simulation–optimization model and the lumped MINLP model—in multireservoir system operation and was able to obtain the optimal hedging operating policy within a reasonable time. With this model, supply can be prioritized for various system demands. Also, the hedging operating policy from the proposed model significantly reduced the magnitude of failures during drought periods.