Unsupervised Basis Function Adaptation for Reinforcement Learning

Edward W. Barker, Charl J. Ras

When using reinforcement learning (RL) algorithms to evaluate a policy it is common, given a large state space, to introduce some form of approximation architecture for the value f…