An emotional model embedded reinforcement learning system

When human does a decision-making, h/she finally does it using the various functions in the brain. He/she also has the ability to learn to improve the decision and get better results than before. Reinforcement learning, one of machine learning methods, is mimicking of learning function of the biological brain´s basal ganglia. In this study, we propose a novel method that combines the conventional reinforcement learning with an emotion model which introduced the concept of biological emotion. Our novel method makes it possible for agent to accomplish complicated tasks which can´t be solved by the conventional reinforcement learning method only. Through computer simulations applying the proposed method to path finding problems, it is verified that the proposed method is more effective comparing with the conventional reinforcement learning method.