Adaptive Intelligent Scheduling for ATM Networks
- R.K. Mehra
- B. Ravichandran
- Richard S. Sutton, Department of Computing Science, University of Alberta
Asynchronous Transfer Mode (ATM) is a fast emerging information technology that promises to provide interoperable multi-media services for commercial and defense applications. Unlike commercial broadband networks which ATM was originally designed for, defense or tactical ATM networks must be able to traverse low-rate transmission links. To allow more flexible and efficient use of this limited bandwidth resource, optimal traffic management in ATM networks is critical. This paper demonstrates the feasibility of developing Self-Learning Adaptive (SLA) scheduling algorithms using Reinforcement Learning (RL). This techniques was applied to simulated data and proved to be more efficient than fixed static scheduling methods.
Citation
R. Mehra, B. Ravichandran, R. Sutton. "Adaptive Intelligent Scheduling for ATM Networks". Yale Workshop on Adaptive and Learning Systems, pp 106-111, January 1996.Keywords: | ATM, feasibility, self-learning adaptive, machine learning |
Category: | In Workshop |
BibTeX
@misc{Mehra+al:YaleWorkshoponAdaptiveandLearningSystems96, author = {R.K. Mehra and B. Ravichandran and Richard S. Sutton}, title = {Adaptive Intelligent Scheduling for ATM Networks}, Pages = {106-111}, booktitle = {}, year = 1996, }Last Updated: May 31, 2007
Submitted by Staurt H. Johnson