Evolving Mathematical Functions with Genetic Algorithms - A Team Project on Artificial Intelligence

- In contrast to most of the senior level Artificial Intelligence (AI) courses where AI is taught in the jargon of theoretical computer science with crowd of complex matrix algebra and differential equations, I demystify the subjects and demonstrate that most fundamental ideas behind AI like Neural Networks, Genetic Algorithms, and Fuzzy Logic, etc, are wonderfully simple and straightforward. I give students team projects as opportunities to be exposed to how intelligent systems are actually built up and implemented. In this paper, we discuss such a project developed by a group of three students for evolving mathematical functions with Genetic Algorithms (GA) in Java. In the project, GA is used to solve a symbolic optimization problem where sequence of instructions collectively forming a candidate math function is evolved to represent a reasonably complicated target math function. Students were very interested and deeply involved in the development of this project since they actually a...

• Type: Conference
• Year: 2006
• Where: FECS
• Authors: Zhiguang Xu, Chris Neasbitt, Jared Summers, Bill Conkling
• Keywords: FECS 2006, FECS 2007, Genetic Algorithms, Level Artificial Intelligence, Math Function