Microsoft PowerPoint - Ag4.ppt [Modalitat compatibilitat]Mario Martin – Autumn 2011 LEARNING IN AGENTS AND MULTIAGENTS SYSTEMS Reinforcement Learning Searching for optimal policies I: Bellman equations and optimal policies Mario Martin Universitat politècnica de Catalunya Dept. LSI Mario Martin – Autumn 2011 LEARNING IN AGENTS AND MULTIAGENTS SYSTEMS How to find optimal policies • Bellman equations for value functions • Evaluation of policies • Properti
