An application of the genetic programming technique to strategy development

  • Authors:

  • Koun-Tem Sun;Yi-Chun Lin;Cheng-Yen Wu;Yueh-Min Huang

  • Affiliations:

  • Department of Information and Learning Technology, National University of Tainan, Taiwan;Department of Engineering Science, National Cheng Kung University, Taiwan;Department of Information and Learning Technology, National University of Tainan, Taiwan;Department of Engineering Science, National Cheng Kung University, Taiwan

  • Venue:
  • Expert Systems with Applications: An International Journal
  • Year:
  • 2009

Quantified Score

Hi-index 12.06

Visualization

Abstract

In this paper, we will apply genetic programming (GP) and co-evolution techniques to develop two strategies: the ghost (attacker) and players (survivors) in the Traffic Light Game (a popular game among children). These two strategies compete against each other. By applying the co-evolution technique alongside GP, each strategy is used as an ''imaginary enemy'' from which evolves (is trained in) another strategy. Based on this co-evolutionary process, these final strategies develop: the ghost can effectively capture the players, but the players can also escape from the ghost, rescue partners, and detour around obstacles. The development of these strategies has achieved phenomenal success. The results encourage us to develop more complex strategies or cooperative models such as human learning models, cooperative robotic models, and self-learning of virtual agents.