How not to be a black-box : evolution and genetic-engineering of high-level behaviours
Ik Soo Lim, Daniël Thalmann · 1999
In spite of many success stories in various domains, Genetic Algorithm and Genetic Programming still suffer from some significant pitfalls. Those evolved programs often lack of some important properties such as robustness, comprehensibility, transparency, modifiability and usability of domain knowledge easily available. We attempt to resolve these problems, at least in evolving high-level behaviours, by adopting a technique of conditions-and-behaviours originally used for minimizing the learning space in reinforcement learning. We experimentally validate the approach on a foraging task.