An Indicator-Based Firefly Algorithm for Many-Objective Optimization

The Firefly Algorithm (FA) has demonstrated remarkable effectiveness in tackling a range of optimization issues. Nonetheless, the research carried out so far has primarily concentrated on addressing single-objective and typical multi-objective optimization problems (MOPs). When dealing with a higher quantity of objectives, such as many-objective optimization problems (MaOPs), FA faces challenges in ensuring ample selection pressure and preserving population diversity. To date, FA has yet to be implemented for tackling MaOPs. In this paper, an indicator-based many-objective FA (namely IBMaOFA) is proposed to challenge MaOPs. Firstly, the indicator \({I}_{\varepsilon +}\) is chosen as a fitness function to evaluate the quality of solutions. Then, a convergence guided search strategy is designed to accelerate the search. To reduce the complexity, a random search model is used. Moreover, a diversity preservation strategy is employed as the environmental selection to maintain diversity. Experimental study is conducted on the DTLZ benchmark. Results show that IBMaOFA can achieve competitive performance when compared with five other state-of-the-art approaches.