Optimizing Multireservoir Operation: Hybrid of Bat Algorithm and Differential Evolution

Abstract

This paper introduces an improved bat algorithm (IBA) with a hybrid mutation strategy to improve its global search ability. In an effort to guide the evolution and reinforce the convergence efficiently, the spatial characteristics of the social and cognitive experience of each bat in the population with the differential evolution (DE) algorithm were developed. More specifically, it has been employed in original bat algorithm (BA) six DE mutation mechanisms, namely the explorative and the exploitative mutation operators. The mutation plays an important role to avoid trapping in a local optimal solution, to ensure the search efficiency of a near global optimal solution, and to increase diversity of population. Also, five unimodal and multimodal benchmark functions were used to test the performance of IBA. The results show that the new bat algorithm performs better than the original bat algorithms for each of the test functions. In addition, IBA could keep the diversity of bats and have a better global search performance. It has been demonstrated that the proposed BA can achieve very low standard deviation for 15 runs of the results. Finally, the proposed method is used to solve two benchmark problems of hydropower operations of multireservoir systems, namely four-reservoir and 10- reservoir systems. The obtained results show that the performance of the proposed method is quite comparable with the results of well-developed traditional linear programming (LP) solvers such as LINGO 8.0, in which for the 15 runs, the best results are as close as 99.9 percent of the global solutions of 308.4 and 1,194.44 for the four-reservoir and 10-reservoir systems, respectively.