Approximate Dynamic Programming via Linear Programming
Abstract
The curse of dimensionality gives rise to prohibitive computational requirements that render infeasible the exact solution of large- scale stochastic control problems. We study an efficient method based on linear programming for approximating solutions to such prob(cid:173) lems. The approach "fits" a linear combination of pre- selected basis functions to the dynamic programming cost- to- go function. We develop bounds on the approximation error and present experi(cid:173) mental results in the domain of queueing network control, providing empirical support for the methodology.
Cite
Text
Farias and Roy. "Approximate Dynamic Programming via Linear Programming." Neural Information Processing Systems, 2001.Markdown
[Farias and Roy. "Approximate Dynamic Programming via Linear Programming." Neural Information Processing Systems, 2001.](https://mlanthology.org/neurips/2001/farias2001neurips-approximate/)BibTeX
@inproceedings{farias2001neurips-approximate,
title = {{Approximate Dynamic Programming via Linear Programming}},
author = {Farias, Daniela and Roy, Benjamin V.},
booktitle = {Neural Information Processing Systems},
year = {2001},
pages = {689-695},
url = {https://mlanthology.org/neurips/2001/farias2001neurips-approximate/}
}