Abstract
Many reinforcement learning approaches can be formulated using the theory of Markov decision processes and the associated method of dynamic programming (DP). The value of this theoretical understanding, however, is tempered by many practical concerns. One important question is whether DP-based approaches that use function approximation rather than lookup tables can avoid catastrophic effects on performance. This note presents a result of Bertsekas (1987) which guarantees that small errors in the approximation of a task's optimal value function cannot produce arbitrarily bad performance when actions are selected by a greedy policy. We derive an upper bound on performance loss that is slightly tighter than that in Bertsekas (1987), and we show the extension of the bound to Q-learning (Watkins, 1989). These results provide a partial theoretical rationale for the approximation of value functions, an issue of great practical importance in reinforcement learning.
Article PDF
Similar content being viewed by others
References
Anderson, C.W. (1986). Learning and Problem Solving with Multilayer Connectionist Systems. PhD thesis, University of Massachusetts, Department of Computer and Information Science, University of Massachusetts, Amherst, MA 01003.
Barto, A.G., Bradtke, S.J., and Singh, S.P. (1991). Real-time learning and control using asynchronous dynamic programming. Technical Report TR-91-57, Department of Computer Science, University of Massachusetts.
Barto, A.G., Sutton, R.S., and Anderson, C.W. (1983). Neuronlike elements that can solve difficult learning control problems. IEEE Transactions on Systems, Man, and Cybernetics, 13 (5), 834–846.
Barto, A.G., Sutton, R.S., and Watkins, C.J.C.H. (1990). Learning and sequential decision making. In M. Gabriel and J. Moore (Eds.), Learning and Computational Neuroscience: Foundations of Adaptive Networks, chapter 13. Cambridge, MA: Bradford Books/MIT Press.
Bertsekas, D.P. (1987). Dynamic programming: Deterministic and stochastic models. Englewood Cliffs, NJ: Prentice Hall.
Bradtke, S.J. (1993). Reinforcement learning applied to linear quadratic regulation. In S.J. Hanson, J.D. Cowan, and C.L. Giles (Eds.), Advances in Neural Information Processing Systems 5, San Mateo, CA. IEEE, Morgan Kaufmann.
Porteus, E. (1971). Some bounds for discounted sequential decision processes. Management Science, 19, 7–11.
Sutton, R.S. (1988). Learning to predict by the methods of temporal differences. Machine Learning, 3, 9–44.
Sutton, R.S. (1990). Integrated architectures for learning, planning, and reacting based on approximating dynamic programming. In B.W. Porter and R.H. Mooney (Eds.), Machine Learning: Proceedings of the Seventh International Conference (ML90), pages 216–224, San Mateo, CA. Morgan Kaufmann.
Tesauro, G. (1992). Practical issues in temporal difference learning. Machine Learning, 8 (3/4), 257–277.
Watkins, C.J.C.H. and Dayan, P. (1992). Q-learning. Machine Learning, 8 (3/4), 279–292.
Watkins, C.J.C.H. (1989). Learning from Delayed Rewards. PhD thesis, King's College, University of Cambridge, Cambridge, England.
Werbos, P.J. (1987). Building and understanding adaptive systems: A statistical/numerical approach to factory automation and brain research. IEEE Transactions on Systems, Man, and Cybernetics, 17 (1), 7–20.
Williams, R.J. and Baird, L.C. (1993). Analysis of some incremental variants of policy iteration: First steps toward understanding actor-critic learning systems. Technical Report NU-CCS-93-11, Northeastern University, College of Computer Science, Boston, MA 02115.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Singh, S.P., Yee, R.C. An Upper Bound on the Loss from Approximate Optimal-Value Functions. Machine Learning 16, 227–233 (1994). https://doi.org/10.1023/A:1022693225949
Issue Date:
DOI: https://doi.org/10.1023/A:1022693225949