Exploration tuned reinforcement function

Abstract The aim of this work is to present a method that helps tune the reinforcement function parameters in a reinforcement learning approach. Since the proposal of neural-based implementations for the reinforcement learning paradigm (which reduced learning time and memory requirements to realistic values) reinforcement functions have become the critical components. Using a particular definition for reinforcement functions (RF), we solve, for a specific case, the so-called exploration versus exploitation dilemma through the careful computation of the RF parameter values. The proposed algorithm computes, during the exploration part of the learning phase, an estimate for the parameter values. Experiments with the mobile robot Nomad 200 validate our proposals.