Decomposition of Reinforcement Learning for Admission Control of Self-Similar Call Arrival Processes

Abstract

This paper presents predictive gain scheduling, a technique for simplify(cid:173) ing reinforcement learning problems by decomposition. Link admission control of self-similar call traffic is used to demonstrate the technique. The control problem is decomposed into on-line prediction of near-fu(cid:173) ture call arrival rates, and precomputation of policies for Poisson call ar(cid:173) rival processes. At decision time, the predictions are used to select among the policies. Simulations show that this technique results in sig(cid:173) nificantly faster learning without any performance loss, compared to a reinforcement learning controller that does not decompose the problem.

Cite

Text

Carlström. "Decomposition of Reinforcement Learning for Admission Control of Self-Similar Call Arrival Processes." Neural Information Processing Systems, 2000.

Markdown

[Carlström. "Decomposition of Reinforcement Learning for Admission Control of Self-Similar Call Arrival Processes." Neural Information Processing Systems, 2000.](https://mlanthology.org/neurips/2000/carlstrom2000neurips-decomposition/)

BibTeX

@inproceedings{carlstrom2000neurips-decomposition,
  title     = {{Decomposition of Reinforcement Learning for Admission Control of Self-Similar Call Arrival Processes}},
  author    = {Carlström, Jakob},
  booktitle = {Neural Information Processing Systems},
  year      = {2000},
  pages     = {1033-1039},
  url       = {https://mlanthology.org/neurips/2000/carlstrom2000neurips-decomposition/}
}