Impact of problem centralization in distributed constraint optimization algorithms
Proceedings of the fourth international joint conference on Autonomous agents and multiagent systems
Completeness and performance of the APO algorithm
Journal of Artificial Intelligence Research
The power of ants in solving Distributed Constraint Satisfaction Problems
Applied Soft Computing
Solving distributed CSPs using dynamic, partial centralization without explicit constraint passing
PRIMA'10 Proceedings of the 13th international conference on Principles and Practice of Multi-Agent Systems
Improving the privacy of the asynchronous partial overlay protocol
Multiagent and Grid Systems - Principles and Practice of Multi-Agent Systems
| Hi-index | 0.00 |
This dissertation focuses on using cooperative mediation to solve distributed problems that are shared between cooperative autonomous agents. Cooperative mediation involves controlled centralization of overlapping subproblems in order to quickly converge on solutions. This technique represents a new paradigm in distributed problem solving that simultaneously exploits the speed of centralized algorithms while using the power of distributed algorithms to identify relevant problem substructures. In this dissertation, we introduce the cooperative mediation paradigm and present three new algorithms used to solve various distributed problems. Amongst these algorithms are asynchronous partial overlay (APO) for solving distributed constraint satisfaction problems, optimal asynchronous partial overlay (OptAPO) for solving distributed constraint optimization problems, and scalable, periodic, anytime mediation ( SPAM) for solving real-time, distributed resource allocation problems. These algorithms have been extensively tested and in the case of APO and OptAPO are shown to be superior to all other current techniques used to solve these problems.