In this paper, to reduce the redundant attractions and incorrect directions of firefly algorithm (FA), a distance-guided selection approach (DSFA) is proposed, which consists of a distance-guided mechanism and selection strategy. Where the designed distance-guided mechanism reduces the attractions and plays as a classifier for global search and local search, the suggested selection strategy can avoid local search falling into traps, thereby increasing the probability of correct direction. With the good cooperation of these two approaches, DSFA obtains a good balance of exploration and exploitation. To confirm the performance of the proposed algorithm, excessive experiments are conducted on CEC2013 benchmark functions, large-scale optimization problems CEC2008, and software defect prediction (SDP). In the comparison with the 5 advanced FA variants, DSFA provides the optimal solutions to most CEC2013 problems. Besides, when facing the problems of class imbalance and the dimensional explosion of datasets, DSFA greatly improves the performance of machine learning classifiers employed by SDP. It can be concluded that DSFA is an effective method for global continuous optimization problems.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.