A Markov Decision Process (MDP) is a natural framework for formulating sequential decision-making problems under uncertainty. In recent years, researchers have greatly advanced algorithms for learning and acting in MDPs. This article reviews such algorithms, beginning with well-known dynamic programming methods for solving MDPs such as policy iteration and value iteration, then describes approximate dynamic programming methods such as trajectory based value iteration, and finally moves to reinforcement learning methods such as Q-Learning, SARSA, and least-squares policy iteration. We describe algorithms in a unified framework, giving pseudocode together with memory and iteration complexity analysis for each. Empirical evaluations of these techniques with four representations across four domains, provide insight into how these algorithms perform with various feature sets in terms of running time and performance.
Thank you for visiting our website. Would you like to provide feedback on how we could improve your experience?
This site does not use any third party cookies with one exception — it uses cookies from Google to deliver its services and to analyze traffic.Learn More.