Modelling of Temperament in an Associative Reinforcement Learning Agent
Zhanna V. Zatuchna, AJ Bagnall · UEA Digital Repository (University of East Anglia) · 2006
The idea of temperament refers to the essential properties of the central nervous system that can produce variations in behaviour and influence the ability of an individual to learn and adapt itself to a complex environment. The research represents an attempt to model certain biological aspects of temperament as alternative learning mechanisms. We investigate the influence of the ‘virtual temperament’ on the effectiveness of the learning in maze environments and evaluate the performance of the learning algorithms on two extensive sets of maze problems.