Multi-Objective Reservoir Operating Strategies by Genetic Algorithm and Nonlinear Programming (GA–NLP) Hybrid Approach

Abstract

This paper presents a genetic algorithm and nonlinear programming (GA–NLP) hybrid model to derive steady state optimal reservoir operating policies for a multi-purpose reservoir. In the present study, the objective is maximizing the net benefits from all the crops in the command area considering yield response to water deficit and the hydro-power generation subject to constraints on reservoir water balance, storage bounds, channel capacities and minimum water requirements. Decision variables of the model are fortnight water allocations to each crop and d/s release for hydro-power generation. The model developed is applied to the Nagarjunasagar multi-purpose reservoir in Andhra Pradesh, India. Reservoir operating policies are developed for three strategies by fixing appropriate weightages to irrigation and power generation in the objective function. Various levels of dependable inflows entering into the reservoir (75% and 80%) are considered in each strategy in the present study. Optimal policies obtained by the proposed model are validated through simulation and compared with standard operating policy (SOP). Reliability of meeting minimum flow requirements with developed policy is found to be superior to SOP. Trade-off curve is developed for the both objectives, and equal priority case is found to be superior yielding maximum total benefit from the project. The results obtained using the proposed model give many alternative strategies for the multi-purpose reservoir managers to take decisions. Results reveal that GA–NLP model can be effectively used for optimal allocation of limited available water resources to any reservoir.