EvoDCMMO: Benchmarking and solving dynamic constrained multimodal optimization problems

doi:10.1016/j.swevo.2022.101184

EvoDCMMO: Benchmarking and solving dynamic constrained multimodal optimization problems

2022-12-01

2

PRE

AI

DOI：10.1016/j.swevo.2022.101184

Original

Original request for help

Save

Abstract

En 中文

In the field of evolutionary computation, many dynamic optimization problems whose fitness values and/or constraints are changed over time have been studied. As the research progresses, complex dynamic optimization problems draw the attention of researchers, such as the dynamic constrained optimization problems with constraints and the dynamic multimodal optimization problems with multiple optima in each environment. Dynamic, constrained, and multimodal properties frequently appear in real-world applications, but the dynamic constrained multimodal optimization problems (DCMMOPs) that simultaneously possess all these properties have not been studied yet. This paper is an extended version of our conference work, aiming to investigate the effect of different techniques for solving DCMMOP. Compared with the previous work, we proposed a more challenging benchmark suite and designed a framework called EvoDCMMO to make further research on DCMMOPs. The new challenge of the proposed benchmark problems arises from the additional constraint based on the fitness values. More techniques, including the optimizers, niching methods, constraint handling techniques and dynamic response strategies are implemented in EvoDCMMO to carry out the comparative studies for solving DCMMOPs and discussed in our work.

Keywords:

Dynamic optimization

Multimodal optimization

Constrained optimization

Evolutionary computation

EvoDCMMO: Benchmarking and solving dynamic constrained multimodal optimization problems

Abstract

Journal

Organization