Computer Science and Engineering
In multi-agent resource allocation, the goal is to allocate resources to agents in a fair and efficient manner based on the preferences, priorities, and endowments of the agents. Immediate applications include scheduling, online barter markets, and allocation of CPU and memory resources in cloud computing. The research project draws on principles and ideas from algorithm design as well as game theory.
This project will involve a survey of classic as well as new resource allocation algorithms. It will also involve an implementation of some of the state of the art algorithms. The student should preferably have taken courses in algorithm design/computational complexity or software engineering, and have interest in game theory.
School
Research Area
Algorithms | Artificial intelligence
- Research Environment
- Expected Outcomes
- Supervisory team
- Reference Material/Links
The project will involve working with members of the Algorithmic Decision Theory group.
This project will require the student to learn about recent developments in resource algorithms. Some algorithms which have not been implemented will be both implemented and tested by the student.
The expected outcome of the project will be a technical report which surveys the algorithms considered during the project as well as an understanding of some of they key issues involved in the field multi-agent resource allocation. One other major outcome will be implementation of some state of the art resource allocation algorithms and evaluation of their relative performance over various metrics.
- Developments in Multi-Agent Fair Allocation
- Y. Chevaleyre, P. E. Dunne, U. Endriss, J. Lang, M. Lemaˆitre, N. Maudet, J. Padget, S. Phelps, J. A. Rodr ´iguez-Aguilar, and P. Sousa. Issues in multiagent resource allocation. Informatica, 30:3–31, 2006.