<p>The optimal operation of reservoirs is a crucial task in water resource management. A robust and simple optimisation technique is required to achieve the best optimal solution. In the present study, a random elites mean hybrid (REMH) algorithm is proposed, which is based on the random elites mean concept while combining the strengths of the Rao-1, Rao-2 and fully informed search (FIS) algorithms. The concept of random elites improves the exploitation capacity of the algorithm, whereas the combination of algorithms enhances the exploration capability of the REMH algorithm, leading to a better chance of finding the global optimal solution. The efficiency of the proposed algorithm is investigated by solving unconstrained mathematical benchmark problems, a multireservoir benchmark problem, and a real-world Dharoi reservoir optimisation problem for 70% dependable inflow condition for multiple years. The proposed optimal solution resulted in a 17% increase in the crop area over the actual cropping pattern in the Dharoi command area. The significance of the proposed REMH algorithm is evaluated using the Friedman test and TOPSIS for unconstrained and engineering optimisation problems, respectively. Statistical analysis revealed that the proposed algorithm outperforms the Rao algorithm and FIS algorithm on mathematical benchmarks and engineering design problems. The proposed algorithm is useful for obtaining the optimal solution of engineering problems, including water resource management.</p>