Lookahead And Latent Learning In ZCS

Learning Classifier Systems use reinforcement learning, evolutionary computing and/or heuristics to develop adaptive systems. This paper extends the ZCS Learning Classifier System to improve its internal modelling capabilities. Initially, results are presented which show performance in a traditional reinforcement learning task incorporating lookahead within the rule structure. Then a mechanism for effective learning without external reward is examined which enables the simple learning system to build a full map of the task. That is, ZCS is shown to learn under a latent learning scenario using the lookahead scheme. Its ability to form maps in reinforcement learning tasks is then considered. keywords: animat, genetic algorithm, internal models, learning classifier system, multiplexer task.