Previous work on machine scheduling has considered the case of agents who control the scheduled jobs and attempt to minimize their own completion time. We argue that in cloud and grid computing settings, different machines cannot be considered to be fully cooperative as they may belong to competing economic entities, and that agents can easily move their jobs between competing providers. We therefore consider a setting in which the machines are also controlled by selfish agents, and attempt to maximize their own gains by strategically selecting their scheduling policy. We analyze the equilibria that arise due to competition in this 2-sided setting. In particular, not only do we require that the jobs will be in equilibrium with one another, but also that the schedulers' policies will be in equilibrium. We also consider different mixtures of classic deterministic scheduling policies and random scheduling policies.
|Original language||American English|
|Title of host publication||Proceedings of the 24th AAAI Conference on Artificial Intelligence, AAAI 2010|
|Number of pages||6|
|State||Published - 15 Jul 2010|
|Event||24th AAAI Conference on Artificial Intelligence, AAAI 2010 - Atlanta, United States|
Duration: 11 Jul 2010 → 15 Jul 2010
|Name||Proceedings of the 24th AAAI Conference on Artificial Intelligence, AAAI 2010|
|Conference||24th AAAI Conference on Artificial Intelligence, AAAI 2010|
|Period||11/07/10 → 15/07/10|
Bibliographical notePublisher Copyright:
© 2010, Association for the Advancement of Artificial Intelligence (www.aaai.org). All rights reserved.