Project Description

In this paper, we provide an exhaustive analysis of the utilization of pair-potential energy functions (PPFs) to generate a reference point set and to increase the diversity properties of Pareto front approximations generated by multi-objective evolutionary algorithms (MOEAs).

According to our experimental results, the utilization of PPF-based mechanisms leads to the generation of high-diversity Pareto front approximations regardless of the underlying geometry.

That is, it is shown that they are promising diversity-preserving mechanisms in multi-objective optimization for algorithm design, algorithm evaluation, and problem analysis.