Abstract
Markov Decision Processes (MDPs) generalize the model underlying classical planning by allowing actions with stochastic effects and fully observable states. In this chapter, we look at a variety of MDP models and the basic algorithms for solving them: from offline methods based on dynamic programming and heuristic search, to online methods where the action to do next is obtained by solving simplifications, like finite-horizon versions of the problem or deterministic relaxations.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
Author information
Authors and Affiliations
Rights and permissions
Copyright information
© 2013 Springer Nature Switzerland AG
About this chapter
Cite this chapter
Geffner, H., Bonet, B. (2013). MDP Planning: Stochastic Actions and Full Feedback. In: A Concise Introduction to Models and Methods for Automated Planning. Synthesis Lectures on Artificial Intelligence and Machine Learning. Springer, Cham. https://doi.org/10.1007/978-3-031-01564-9_6
Download citation
DOI: https://doi.org/10.1007/978-3-031-01564-9_6
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-00436-0
Online ISBN: 978-3-031-01564-9
eBook Packages: Synthesis Collection of Technology (R0)eBColl Synthesis Collection 5