Multi-agent Coordination through Mutualistic Interactions
In this paper we present an ecologically-inspired approach to agent coordination. Mutualistic networks of interacting species in nature possess characteristics that provide the systems they represent with features of stability, minimised competition, and increased biodiversity. We take inspiration from some of the ecological mechanisms that operate at the interaction level in mutualistic interactions, and which are believed to be responsible for the emergence of these system level patterns, in order to promote this structural organisation in networks of interacting agents, enhancing in this way their cooperative abilities. We demonstrate that given plausible starting conditions, we can expect mutualistic features to appear in self-organising agent systems, and we compare them with natural ones to show how the characteristics displayed by ecologically inspired networks of agents are similar to those found in natural communities. We argue that the presence of these patterns in agent interaction networks confer these systems with properties similar to those found in mutualistic communities found in the real world.
Keywordsagents coordination ecologically-inspired interactions complex systems mutualism emergent behaviour cooperation
Unable to display preview. Download preview PDF.
- 8.de Pinnick Bas, A.P., Dupplaw, D., Kotoulas, S., Siebes, R.: The openknowledge kernel. International Journal of Applied Mathematics and Computer Sciences 4(3), 162–167 (2007)Google Scholar
- 9.Joppa, L.N., Montoya, J.M., Solé, R.V., Sanderson, J., Pimm, S.L.: On nestedness in ecological networks. Evolutionary Ecology Research 12(1), 35–46 (2010)Google Scholar
- 10.Lurgi, M.: Ecologically inspired agent networks. Master’s thesis, University of Edinburgh. College of Science and Engineering. School of Informatics (2010)Google Scholar
- 19.Villalba, C., Zambonelli, F.: An nature-inspired approach for large-scale pervasive service ecosystems. In: Proceedings of the 4th AAMAS Workshop on Massive Multiagent Systems. IFAAMAS (May 2009)Google Scholar