Adaptive Fitness Predictors in Coevolutionary Cartesian Genetic Programming

Abstract

In genetic programming (GP), computer programs are often coevolved with training data subsets that are known as fitness predictors. In order to maximize performance of GP, it is important to find the most suitable parameters of coevolution, particularly the fitness predictor size. This is a very time-consuming process as the predictor size depends on a given application, and many experiments have to be performed to find its suitable size. A new method is proposed which enables us to automatically adapt the predictor and its size for a given problem and thus to reduce not only the time of evolution, but also the time needed to tune the evolutionary algorithm. The method was implemented in the context of Cartesian genetic programming and evaluated using five symbolic regression problems and three image filter design problems. In comparison with three different CGP implementations, the time required by CGP search was reduced while the quality of results remained unaffected.

Cartesian genetic programming: its status and future

“…Šikulová et al [14,90] developed a combination of fitness prediction with coevolution in CGP to reduce the number of expensive full fitness evaluations. The method replaces some of the objective fitness evaluations with an alternative fitness calculated using a fitness predictor.…”

“…They also use two archives one of fitness trainers and one containing the best evolved fitness predictors. The archive [14]. Some of the objective fitness evaluations are replaced with an alternative fitness calculated using a fitness predictor.…”

“…Drahosova et al [14] evaluated their method on suites of varied symbolic regression problems and image filter design problems. They showed that the coevolutionary approach outperforms standard CGP in terms of CPU time required to converge.…”