Methods for acceleration of learning process of Reinforcement Learning Neuro-Fuzzy Hierarchical Politree model

This paper presents two methods for accelerating the learning process of Reinforcement Learning - Neuro-Fuzzy Hierarchical Politree model (RL-NFHP): policy Q-DC-Roulette and early stopping. This model is used to provide an agent with intelligence, making it autonomous, due to the capacity of ratiocinate (infer actions) and learning, acquired knowledge through interaction with the environment. The characteristics of the RL-NFHP allow the agent to learn automatically its structure and action for each state. The RL-NFHP model was evaluated in an application benchmark known in the area of autonomous agents: car mountain problem. The results demonstrate the acceleration of learning process and the potential of this model, which works without any prior information, such as number of rules, rules of specification, or number of partitions that the input space should possess.