Development of the marine predators algorithm for optimizing the performance of water supply reservoirs

Optimal reservoir operation involves complex decision making. The marine predators algorithm (MPA) is herein applied and evaluated with several mathematical functions and with the optimized operation of the Aydogmush reservoir (East Azerbaijan province, Iran) that minimizes the deficit of agricultural water supplied to downstream lands. Reservoir operation covers a baseline period (1987–2000) and a period of climate change (2026–2039). The MPA’s reservoir operation results are compared those obtained with the genetic algorithm (GA). The calculated operating policies are evaluated based on indexes of reliability, resiliency and vulnerability. A comparison of reservoir water releases, water supply deficit and reservoir storage from five runs under baseline and climate change periods indicates better performance of the MPA-calculated reservoir operation than the GAs in meeting downstream water demand. The efficiency indexes show that, for example, the reliability of the operating policy obtained with the MPA is larger by 54 and 34% compared to the GAs in the baseline and the climate change periods, respectively.