Approximate Equivalence of Markov Decision Processes


We consider the problem of finding the minimal ǫ-equivalent MDP for an MDP given in its tabular form. We show that the problem is NP-Hard and then give a bicriteria approximation algorithm to the problem. We suggest that the right measure for finding minimal ǫ-equivalent model is L1 rather than L∞ by giving both an example, which demonstrates the drawback of using L∞, and performance guarantees for using L1. In addition, we give a polynomial algorithm that decides whether two MDPs are equivalent.

DOI: 10.1007/978-3-540-45167-9_42

Extracted Key Phrases

2 Figures and Tables

Cite this paper

@inproceedings{EvenDar2003ApproximateEO, title={Approximate Equivalence of Markov Decision Processes}, author={Eyal Even-Dar and Yishay Mansour}, booktitle={COLT}, year={2003} }