A Computational Model of Habit Learning to Enable Ambient Support for Lifestyle Change

Agent-based applications have the potential to assist humans in their lifestyle change, for instance eliminating addictive behaviours or adopting new healthy behaviours. In order to provide adequate support, agents should take into consideration the main mechanisms underlying behaviour formation and change. Within this process habits play a crucial role: automatic behaviours that are developed unconsciously and may persist without the presence of any goals. Inspired by elements from neurological literature, a computational model of habit formation and change was developed as a basis for support agents able to assist humans in lifestyle and behaviour change. Simulations are presented showing that the model exhibits realistic human-like behaviour. © 2011 Springer-Verlag