Evolutionary learning in agent-based combat simulation

  • Authors:
  • Tomonari Honda;Hiroshi Sato;Akira Namatame

  • Affiliations:
  • Dept of Computer Science, National Defense Academy, Yokosuka, Japan;Dept of Computer Science, National Defense Academy, Yokosuka, Japan;Dept of Computer Science, National Defense Academy, Yokosuka, Japan

  • Venue:
  • ACRI'06 Proceedings of the 7th international conference on Cellular Automata for Research and Industry
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this paper, we consider one of old-age problems about trade-off relation between homogeneity and diversity We investigate combat based on agent-based simulation, not conventional mathematical model based on attrition By introducing synthetic approach and adapting evolutionary learning to action rules that are expressed by a combination of parameters in combat simulation, we focus on the interaction between sets of action rules For searching how many sets of action rules does work well, we change the number of sets of action rules And we make statistical analysis and show that there is good intermediate stage between high homogeneity and high diversity in group.