AN MOEA-BASED METHOD TO TUNE EA PARAMETERS ON MULTIPLE OBJECTIVE FUNCTIONS

In this paper, we demonstrate the benefits of using a multi-objective approach when tuning the parameters of an Evolutionary Algorithm. To overcome the specific challenges that arise when using a meta-algorithm for parameter tuning on multiple functions, we introduce a new algorithm called the Multi-Function Evolutionary Tuning Algorithm (M-FETA) that is able to approximate the parameter Pareto front effectively. The results of the experiments illustrate how the approximated Parameter Pareto front can be used to gain insights, identify ‘generalists’, and study the robustness of the algorithm to be tuned.