Lion swarm optimization algorithm for comparative study with application to optimal dispatch of cascade hydropower stations

Lion swarm optimization (LSO) algorithm that based on the natural division of labor among lion king, lionesses and lion cubs in a pack of lions is recently introduced. To evaluate the exploration and the exploitation of the LSO algorithm comprehensively, an intensive study based on optimization problems is necessary. In this work, we firstly present the revised version of the LSO algorithm in detail. Secondly, the efficiency of LSO is evaluating using quantitative analysis, convergence analysis, statistical analysis, and robustness analysis on 60 classical numerical test problems, encompassing the Uni-modal, the Multi-modal, the Separable, the Non-separable, and the Multi-dimension problems. For comparison purposes, the results obtained by the LSO algorithm are compared against a large set of state-of-the-art optimization methods. The comparative results show that the LSO can provide significantly superior results for the US, the UN, and the MS problems regarding convergence speed, robustness, success rate, time complexity, and optimization accuracy compared with the other optimizers, and present very competitive results in terms of those indicators compared with the other optimizers. Finally, to check the applicability and robustness of the LSO algorithm, a case study on optimal dispatch problem of China's Wujiang cascade hydropower stations shows that the LSO can obtain well and reliable optimal results with average generation of 122.421180 10(8) kWh, 103.463636 10 8 kWh, and 99.3826340 10(8) kWh for three different scenarios (i.e. the wet year, the normal year and the dry year), which are satisfying compared with that of the GA, the improved CS, and the PSO in terms of optimization accuracy. Besides, regarding the convergence speed, the results are also competitive. Therefore, we can conclude that the LSO is an efficient method for solving complex problems with correlative decision variables with simple structure and excellent convergence speed. (C) 2019 Elsevier B.V. All rights reserved.