随机环境下惩罚和奖励对任务型团队演化的影响  

The Impact of Punishment and Reward on the Evolution of Task-Oriented Teams in Random Environment

作　　者：平安 张悟移[1] 杜春澎 何治学 韩正涛 PING An;ZHANG Wuyi;DU Chunpeng;HE Zhixue;HAN Zhengtao(Faculty of Management and Economics,Kunming University of Science and Technology,Kunming 650093;School of Statistics and Mathematics,Yunnan University of Finance and Economics,Kunming 650221)

机构地区：[1]昆明理工大学管理与经济学院,昆明650093 [2]云南财经大学统计与数学学院,昆明650221

出　　处：《系统科学与数学》2023年第10期2503-2524,共22页Journal of Systems Science and Mathematical Sciences

基　　金：国家自然科学基金(11931015,7216403);中国工程院院地合作重大咨询项目(2020YNZH)资助课题。

摘　　要：任务型团队的工作模式是社会合作及生产中重要的一环.但随机噪声环境对团队组建的影响目前还鲜有研究.文章基于演化博弈理论和强化学习算法,构建智能体团队模型,研究噪声环境下惩罚和奖励机制对于任务型团队演化的影响.研究发现外部环境的随机程度影响团队的形成,在特定的环境下使用合适的奖励和惩罚能够促进个体的合作意愿.其中奖励主要影响团队的形成,惩罚对合作行为促进更大.Task-oriented teamwork is a fundamental aspect of social cooperation and production.However,there has been little research on the influence of random noise environment on team formation.To investigate the influence of punishment and reward mechanisms on the development of task-oriented teams in noisy environment,this research proposes an agent task-based team model based on evolutionary game theory and reinforcement learning algorithm.The results find that the randomness of the external environment impacts team formation.Appropriate punishment and reward for specific environment could improve individuals'cooperative willingness.Reward effects team formation mainly.Punishment enhances cooperation more.

关 键 词：不确定性 公共物品博弈 惩罚 奖励 多智能体强化学习 团队形成 合作演化 

分 类 号：O225[理学—运筹学与控制论] TP18[理学—数学]