Computer systems are rapidly becoming so complex that maintaining them with human support staffs will be prohibitively expensive and inefficient. In response, visionaries have begun proposing that computer systems be imbued with the ability to configure themselves, diagnose failures, and ultimately repair themselves in response to these failures. However, despite convincing arguments that such a shift would be desirable, as of yet there has been little concrete progress made towards this goal. We view these problems as fundamentally emphmachine learning challenges. Hence, this article presents a new network simulator designed to study the application of machine learning methods from a system-wide perspective. We also introduce learning-based methods for addressing the problems of job routing and CPU scheduling in the networks we simulate. Our experimental results verify that methods using machine learning outperform reasonable heuristic and hand-coded approaches on example networks designed to capture many of the complexities that exist in real systems.

PDF, PS, HTML

Engineering Applications of Artificial Intelligence, 17(7)(7):855-869, October 2004. Corrected version.

@Article{whiteson:eaai04, title={Adaptive Job Routing and Scheduling}, author={Shimon Whiteson and Peter Stone}, volume={17(7)}, journal={Engineering Applications of Artificial Intelligence}, number={7}, month={October}, pages={855-869}, note={Corrected version}, url="http://nn.cs.utexas.edu/?whiteson:eaai04", year={2004} }