An Exploration Method for Constructing Semi-Markov Decision Process Models Using k-Certainty Exploration Method and Fuzz-ART for Continuous State Space Environments

Abstract

k-certainty exploration method is a very efficient method to construct Markov Decision Process models of unknown environments. However, it is impossible to apply this method to environments which include continuous state space. This paper describes a new exploration method using k-certainty exploration method and Fuzzy-ART to represent these environments as Semi-Markov Decision Process models.