计算机与现代化

• 计算机仿真 • 上一篇    下一篇

 基于Q-学习和行为树的CGF空战行为决策

  

  1. 海军航空工程学院信息融合研究所,山东烟台264001
  • 收稿日期:2017-01-20 出版日期:2017-05-26 发布日期:2017-05-31
  • 作者简介: 方君(1979-),男,安徽怀宁人,海军航空工程学院信息融合研究所讲师,硕士,研究方向:作战仿真,飞行仿真。
  • 基金资助:
     国家自然科学基金重大研究计划(91538201); 泰山学者专项基金资助项目(ts201511020)

 Air Bat Strategies of CGF Based on Q-learning and Behavior Tree

  1. Institute of Information Fusion, Naval Aeronautical and Astronautical University, Yantai 264001, China
  • Received:2017-01-20 Online:2017-05-26 Published:2017-05-31

摘要:  空战行为决策的智能性是学术界关注的重要问题之一。提出一种基于Q-学习和行为树的CGF空战行为决策方法。通过构建CGF空战行为树模型,实现CGF智能行为;通过在行为树上的Q-学习,使CGF具有不断进化的能力。仿真结果表明,该算法在与传统算法对抗中,性能优势明显且学习能力较强。

关键词: 空战决策, 人工智能, 行为树, Q-学习

Abstract:  The intelligence of air bat strategies is one of the important problems. A new method for air bat strategies of CGF was proposed based on Q-learning and behavior tree. The intelligence of CGF was formed through establishing behavior tree. And through Q-learning on behavior tree, the evolutionary ability was gained for CGF. Simulation shows that the method performs better and with a stronger learning ability when it combats with traditional algorithm.

Key words:  air bat strategies, artificial intelligence, behavior tree, Q-learning