Resource allocation in the grid using reinforcement learning

By Aram Galstyan, Karl Czajkowski, Kristina Lerman. 2004

135 citations recorded by the source corpus.

In this paper we study a minimalist decentralized algorithm for resource allocation in a simplified Grid-like environment. We consider a system consisting of large number of heterogenous reinforcement learning agents that share common resources for their computational needs. There is no communication between the agents: the only information that agents receive is the (expected) completion time of a job it submitted to a particular resource and which serves as a reinforcement signal for the agent. The results of our experiments suggest that reinforcement learning can be used to improve the quality of resource allocation in large scale heterogenous system.

Subfields: Artificial Intelligence, Computer Networks and Communications, Electrical and Electronic Engineering

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Networks and Communications

In Electrical and Electronic Engineering

Citations in the corpus

A Multi-Agent Framework for Packet Routing in Wireless Sensor Networks (2015, methodology): “[34] applied a Q-learning algorithm with -greedy exploration to develop a decentralised resource allocation mechanism; Gomes and Kowalczyk [35]”
Learning the IPA market with individual and social rewards (2007, unknown): “In Galstyan et al. [14], agents learn which resource nodes to submit their jobs to, given that the nodes are managed by local schedulers.”
Brain Meets Brawn: Why Grid and Agents Need Each Other (2004, background): “As an early example of such a tighter coupling, we point to work on agent-based resource selection, in which re-enforcement-based learning is used to drive the assignment of tasks to resources [16].”
Dynamic analysis of multiagent Q-learning with ε-greedy exploration (2009, methodology): “For example: Galstyan et al. (2004) applies the algorithm to develop a decentralized resource allocation mechanism; Gomes and Kowalczyk (2007) study the problem of learning demand functions; and Ziogos et al.”
A Taxonomy of Workflow Management Systems for Grid Computing (2005, background): “ts. The on-line learning approach predicts task execution performance from on-line experience without prior knowledge of the environment’s dynamics. For example, Buyya et al. [22] and Galstyan et al. [56] maps a job onto a ‘best’ Grid resource by learning the completion time of most recent jobs subm”
Applying reinforcement learning towards automating resource allocation and application scalability in the cloud (2013, methodology): “[23] implemented a decentralised multi-agent reinforcement learning approach to Grid resource allocation.”