Evolutionary algorithm-based multiobjective reservoir operation policy optimisation under uncertainty

Abstract Reservoir operation optimisation is a decision support tool to assist reservoir operators with water release decisions achieve management objectives, such as maximising supply security, mitigating flood risk, and hydroelectric power generation. The effectiveness of subject uncertainty in system inputs, inflow therefore, methods stochastic dynamic programming (SDP) have been traditionally used. However, these suffer from the three curses dimensionality, modelling, multiple objectives. Evolutionary algorithm (EA)-based simulation-optimisation frameworks Multi-Objective Direct Policy Search offer new paradigm for multiobjective under uncertainty, directly addressing shortcomings SDP-based methods. They also enable consideration input represented using ensemble forecasts that become more accessible recently. there no universally agreed approach incorporate into EA-based policy it not clear which effective. Therefore, this study conducts comparative analysis demonstrate advantages limitations different approaches account via real-world case study; provide guidance on selection appropriate approaches. Based results obtained, evident each has both limitations. A suitable needs be carefully selected based study, e.g., whether hard constraint required, or well-established decision-making process exists. In addition, potential gaps future research are identified.