A Decentralized Multi-agent Algorithm for the Set Partitioning Problem
A decentralized algorithm for solving set partitioning problems (SPP) has numerous applications in multi-agent systems. Apart from its relevance for problems of operations research, the SPP is equivalent to clustering of agents as well as the generation of coalition structures. SPADA, the algorithm proposed in this paper, does not use central metrics, relies exclusively on local agent knowledge, and respects agent autonomy. By trying to increase its own benefit, each subset of agents, representing a coalition or cluster, contributes to establishing a suitable partitioning. Agents are at liberty to decide if they want to join or leave a subset. All operations that change the composition of these subsets or the acquaintances of agents can be mapped onto a graph that represents these relations. The algorithm is evaluated in a scenario of decentralized energy management where it is used as a self-organization mechanism. The evaluations show that the quality of the partitioning is within 10% of the solutions found by a particle swarm optimizer with global knowledge.
KeywordsSelf-Organization Distributed Problem Solving Set Partitioning Problem Coalition Formation Clustering
Unable to display preview. Download preview PDF.
- 1.Repast Simphony, http://repast.sourceforge.net (accessed June 1, 2012)
- 2.Abdallah, S., Lesser, V.: Organization-Based Cooperative Coalition Formation. In: Int. Conference on Intelligent Agent Technology, pp. 162–168 (2004)Google Scholar
- 3.Al Faruque, M.A., Krist, R., Henkel, J.: ADAM: run-time agent-based distributed application mapping for on-chip communication. In: Proc. of the 45th Annual Design Automation Conference, pp. 760–765. ACM (2008)Google Scholar
- 4.Anders, G., Seebach, H., Nafz, F., Steghöfer, J.-P., Reif, W.: Decentralized Reconfiguration for Self-Organizing Resource-Flow Systems Based on Local Knowledge. In: 8th IEEE Int. Conference and Workshops on Engineering of Autonomic and Autonomous Systems (EASe), pp. 20–31 (2011)Google Scholar
- 5.Anders, G., Siefert, F., Steghöfer, J.-P., Seebach, H., Nafz, F., Reif, W.: Structuring and Controlling Distributed Power Sources by Autonomous Virtual Power Plants. In: Proc. of the Power & Energy Student Summit, pp. 40–42 (October 2010)Google Scholar
- 9.Kennedy, J., Eberhart, R.: Particle Swarm Optimization. In: Proc. of the IEEE Int. Conference on Neural Networks, vol. 4, pp. 1942–1948 (1995)Google Scholar
- 10.Motee, N., Sayyar-Rodsari, B.: Optimal partitioning in distributed model predictive control. In: Proc. of the 2003 American Control Conference, vol. 6, pp. 5300–5305 (June 2003)Google Scholar
- 11.Ogston, E., Overeinder, B., Steen, M.V., Brazier, F.: A Method for Decentralized Clustering in Large Multi-Agent Systems. In: Proc. of the 2nd Int. Joint Conference on Autonomous Agents and Multiagent Systems, pp. 789–796 (2003)Google Scholar