Combining Reinforcement Learning with an Emotion based Architecture

P. Vale and L. Custódio (Portugal)


Reinforcement Learning, Emotions, Learning


In this paper an Emotion-Based Agent Architecture used together with a Reinforcement Learning algorithm is described. In order to experiment this architecture, a very simple environment was created and the architecture was compared both to an emotional agent and a Reinforcement Learning agent. The results obtained offer some evidence that the combination of both approaches allows more ”intelligent” behaviors and faster learning.

Important Links:

Go Back