Improved speciation-based Firefly Algorithm in dynamic and uncertain environment

Many real-world optimization problems are dynamic in nature. The applied algorithm in this environment can pose serious challenges, especially when the search space is multimodal with multiple, time-varying optima. To address these challenges, this paper investigates a speciation-based firefly algorithm to enhance the population diversity with aim of generating several populations in different areas in the landscape without knowing the number of optima in it. To improve the performance of the algorithm, multiple adaptation techniques have been used: automatically adapting the number of subswarm in environment with speciation-based firefly algorithm, adapting number of fireflies in each subswarm and also adapting number of active fireflies during the run based on the domain knowledge of fitness landscape. An experimental study is presented with a well-known Moving Peaks Benchmark (MPB) problem to evaluate the performance of the proposed algorithm with other leading algorithms in dynamic environments. The results show the efficiency of the proposed algorithm in comparison with other state-of-the-art algorithms on different configurations and scenarios of this benchmark.