Classifier systems (CS) are used as control architectures for simulated animals or robots in order to decide what to do at each time. We will explain why these systems are good candidates for action selection mechanisms of Non Player Characters. After having described different classifier systems, we will introduce a new CS architecture, acting in a multi-agent environment, which is adapted to the specific constraints of the `Massively Multi-players Online Role Playing Games'.

This paper introduces generalized state values to the anticipatory learning classifier system ACS2. Previous studies with ACS2 showed that the system reliably evolves a generalized predictive model in typical Markov decision process (MDP). The predictive model approximates the state transition function of the MDP in a compact, generalized form. However, it was also shown that the evolving predictive model might be over-general for an accurate representation of reinforcement values. Thus, a function approximation module is added that approximates state values. In combination, actual action choice depends on state values predicted by the means of the predictive model yielding anticipatory behavior. It is shown that the function approximation module accurately generalizes the state value function in the investigated MDP. We also

On behalf of: