Conference Proceeding Article
While Markov Decision Processes (MDPs) have been shown to be effective models for planning under uncertainty, theobjective to minimize the expected cumulative cost is inappropriate for high-stake planning problems. As such, Yu, Lin, and Yan (1998) introduced the Risk-Sensitive MDP (RSMDP) model, where the objective is to find a policy that maximizes the probability that the cumulative cost is within some user-defined cost threshold. In this paper, we revisit this problem and introduce new algorithms that are based on classical techniques, such as depth-first search and dynamic programming, and a recently introduced technique called Topological Value Iteration (TVI). We demonstrate the applicability of our approach on randomly generated MDPs as well as domains from the ICAPS 2011 International Probabilistic Planning Competition (IPPC).
Artificial Intelligence and Robotics | Operations Research, Systems Engineering and Industrial Engineering
Intelligent Systems and Decision Analytics
Proceedings of the Twenty-Fourth International Conference on Automated Planning and Scheduling (ICAPS-14): 21-26 June 2014, Portsmouth, New Hampshire
City or Country
Menlo Park, CA
HOU, Ping; YEOH, William; and VARAKANTHAM, Pradeep Reddy.
Revisiting Risk-Sensitive MDPs: New Algorithms and Results. (2014). Proceedings of the Twenty-Fourth International Conference on Automated Planning and Scheduling (ICAPS-14): 21-26 June 2014, Portsmouth, New Hampshire. 136-144. Research Collection School Of Information Systems.
Available at: http://ink.library.smu.edu.sg/sis_research/2089
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.