Title :
Variable interaction network based on activation and inhibition
Author :
Sekiyama, Kosuke ; Okade, Yukihisa
Author_Institution :
Dept. of Human & Artificial Intelligence Syst., Fukui Univ., Japan
Abstract :
This paper deals with an evolutionary graph network, which represents cooperative and competitive form in the multi-agent systems. We focus on the functional relationship or connectivity of the graph structure, rather than the physical configuration of itself. Each node corresponding to the agent has a variable number of links to the neighborhood nodes, dependent on the internal state and local interactions characterized as activation and inhibition. A simple collective game is considered to exemplify that evolutionary mechanisms of reconfiguration of the graph structure enhances cooperation form.
Keywords :
evolutionary computation; game theory; graph theory; multi-agent systems; activation; competitive form; cooperative form; evolutionary graph network; evolutionary mechanism; graph structure; inhibition; internal state; multiagent system; variable interaction network; Artificial intelligence; Bonding; Competitive intelligence; Glass; Humans; Hydrogen; Intelligent networks; Intelligent systems; Multiagent systems; Teamwork;
Conference_Titel :
Computational Intelligence in Robotics and Automation, 2003. Proceedings. 2003 IEEE International Symposium on
Print_ISBN :
0-7803-7866-0
DOI :
10.1109/CIRA.2003.1222228