基于成员-任务互动的团队有效性多智能体模拟  

Multi-agent Simulation of Team Effectiveness Based on Member-task Interaction

作　　者：董升平[1] 胡斌[1] 张金隆[1] 

机构地区：[1]华中科技大学管理学院,湖北武汉430074

出　　处：《中国管理科学》2008年第5期171-181,共11页Chinese Journal of Management Science

基　　金：国家自然科学基金资助项目(70671048)

摘　　要：团队任务过程和团队人群行为之间通过成员能力与任务要求的匹配,成员在任务处理中学习而能力得以提高等人与任务间的联系而发生成员-任务互动。为分析这种互动对团队有效性的影响,在分析了团队任务过程中各影响因素间的因果关系之后,构建了一个多智能体模拟模型,采用Repast.J 3.1实现了该模型,开发了基于成员-任务互动的团队有效性多智能体模拟系统,并设计了多个实验方案对系统进行了验证和确认,之后该系统被用于分析任务规模和波动幅度的影响。模拟结果表明在团队任务过程中,无论任务规模或关系波动幅度多大,团队成员关系度与任务成本不可能同时达到最优,同时就如何有效促进团队有效性给出了建议。The interactions between the task process and members＇ behavior take place through the match of members＇ ability and demands required by tasks and the enhancement of one＇s ability through learning in task process. To address how the interaction affects team effectiveness, a multi--agent model is put for- ward based on casual anaIyses of various factors involved in team＇s task process. The model is implemen- ted into Member--task Interaction based on Team Effectiveness Simulation System by Repast. J 3.1. Ex- perimental scenarios are designed and simulated to verify and validate the system. Then, we use the sys- tem to explore the effect of task size and fluctuation magnitude of member relations on team effectiveness and task allocation. The results show degree of that team＇s member relationship and task cost can not achieve optimal values simultaneously, and suggestions are provided on how to enhance team effectiveness.

关 键 词：团队有效性 成员一任务互动 多智能体模拟 REPAST 

分 类 号：TP319.9[自动化与计算机技术—计算机软件与理论]