Between MDPs and Semi-MDPs:Learning, Planning, and Representing Knowledge at Multiple Temporal Scales

Authors

Richard S. Sutton, University of Massachusetts - Amherst

Publication Date

1998

Abstract

Learning, planning, and representing knowledge at multiple levels of temporal abstraction are key challenges for AI. In this paper we develop an approach to these problems based on the mathematical framework of reinforcement learning and Markov decisions processes (MDPs). We extend the usual notion of action to include options--whole courses of behavior that may be temporally extended, stochastic, and contingent on events.

Comments

This paper was harvested from CiteSeer

Recommended Citation

Sutton, Richard S., "Between MDPs and Semi-MDPs:Learning, Planning, and Representing Knowledge at Multiple Temporal Scales" (1998). Computer Science Department Faculty Publication Series. 213.
Retrieved from https://scholarworks.umass.edu/cs_faculty_pubs/213