Hybrid of Rule-based Systems Using Genetic Algorithm to Improve Platform Game Performance

摘要

Several attempts in the field for various games have been made using Genetic Algorithm. “Geometry Friends” is one of many platform games. This paper shows performance improvement of the platform game by applying GA. We attempt to determine the parameters for rule-based systems in the “Geometry Friends” game. We perform experiments for the problem not to be solved based on rule-based systems. We divide experimental cases and apply GA to each case. As a result, we could get a set of optimal parameters. Parameters tuning of rule-based systems using GA is more effective to improve performance than rule-based systems.