Application of Sub‐Population Scheduling Algorithm in Multi‐Population Evolutionary Dynamic Optimization
Javidan Kazemi Kordestani, Mohammad Reza Meybodi · 2020
The multi-population approach has been widely used to address dynamic optimization problems (DOPs) by the evolutionary computation (EC) community. Several methods have been already developed to enhance the performance of multi-population algorithms for DOPs. In this chapter we focus on managing function evaluations (FEs) for intelligent distribution of FEs among sub-populations. To this end, we investigate the application of sub-population scheduling (SPS) for improving the performance of multi-population methods when solving DOPs. Eight SPS methods are reviewed, of which two are proposed for the first time in this chapter. Experiments are carried out to investigate these methods and their performance is compared and analyzed. The experimental results show the efficiency of the SPS for improving the multi-population algorithms in dynamic environments.