Multi-mechanism ensemble interior search algorithm to derive optimal hedging rule curves in multi-reservoir systems

Abstract

The operating rules, especially the hedging rules, which are known as useful tools for effective utilization of water from reservoir systems, are commonly applied for managing reservoir systems in severe drought periods. Extracting optimal operation policies for multi-reservoir systems is a challenging issue due to their non-linear, large-scale, and stochastic nature. Therefore, improving optimization techniques is a worthy effort in developing the operating rules for multi-reservoir systems. Hence, this research aimed to develop an efficient optimization method called multi-mechanism ensemble interior search algorithm (MEISA) to tackle various problems in the optimal management of reservoir systems. The MEISA promotes the two main operators of the interior search algorithm (ISA) (i.e., the composition and the mirror operators) to enhance the capability of global and local search. Besides, the proposed algorithm employs an enhanced opposition-based learning (EOBL) operator to improve the best solution and increase the convergence speed. The MEISA also uses an adaptive parameter to balance the global and local search in the proposed algorithm. The results of several test functions demonstrate that the MEISA outperforms other optimizers in terms of the solution accuracy, the quality, and the convergence rate. Finally, the results of optimizing the hedging rules for two real-world single- and multi-reservoir systems in Iran indicate that the MEISA is more suitable efficient in ameliorating irreversible impacts of severe drought periods as compared with other optimization methods. The results show that the MEISA has highly convincing performance in obtaining the optimal hedging rules for the efficient management of multi-reservoir systems.