A model to measure the performance of evolutionary algorithms applied to solve the root identification problem

The existence of a model for representing the performance of metaheuristics applied to solve problems with high computational requirements is paramount to determine the solution quality given a certain avaliable run-time and vice versa. In this work, a statistical model is proposed to describe the performance of evolutionary algorithms applied to solve the Root Identification Problem. Given an unknown problem size, a parameter setting and a performance model are estimated for two well-known evolutionary algorithms, Population-Based Incremental Learning (PBIL) and Cross generational elitist selection Heterogeneus recombination and Cataclismic mutation (CHC). The performance model is validated over a benchmark corresponding to huge search spaces.