Exploitation of User’s Preferences in Reinforcement Learning Decision Support Systems

Contents

A system called COLMAS (COordination Learning in Multi-Agent System) has been developed to investigate how the integration of realistic geosimulation and reinforcement learning might support a decision-maker in the context of cooperative patrolling. COLMAS is a model-driven automated decision support system combining geosimulation and reinforcement learning to compute near optimal solutions. Building upon this hybrid approach, this paper proposes an extended framework to constructively incorporate user preferences providing mixed-initiative generation of further trusted and validated solutions. The proposed approach integrates the user’s preferences in COLMAS by automatically extracting user’s preferred solution.