Title :
Switching dynamics of multi-agent systems: soap froths paradigm
Author :
Lor, W.K.F. ; Szeto, K.Y.
Author_Institution :
Dept. of Electr. & Electron. Eng., Imperial Coll. of Sci., Technol. & Med., London, UK
Abstract :
Monte Carlo simulation of the color switching dynamics in performed on cellular networks using Voronoi construction of point patterns. Association of cells to a particular choice of color is interpreted as representative agents for a particular company. The evolution of color patterns on static cellular network is interpreted as a manifestation of the competition and cooperation between companies. The soap froths paradigm is employed to analyze the stability of the color pattern of the multiagent system. We find existence of phases corresponding to dominant color, as function of the noise level as well as the strength ratio of different color interaction. We observe two phase transitions where the cellular network of agents first transits from a disordered phase into an ordered phase, and then back to a disordered phase as the noise level is increased
Keywords :
Monte Carlo methods; computational geometry; multi-agent systems; neural nets; switching functions; Monte Carlo simulation; Voronoi construction; cellular networks; color pattern stability; color switching dynamics; disordered phase; multiagent system; neural nets; ordered phase; phase transitions; point patterns; soap froths paradigm; switching dynamics; Colored noise; Intelligent systems; Land mobile radio cellular systems; Lattices; Multiagent systems; Network topology; Noise level; Simple object access protocol; Stability; Steady-state;
Conference_Titel :
Neural Networks, 2000. IJCNN 2000, Proceedings of the IEEE-INNS-ENNS International Joint Conference on
Conference_Location :
Como
Print_ISBN :
0-7695-0619-4
DOI :
10.1109/IJCNN.2000.859465
