• 论文 •    

基于多Agent的高层体系结构分布式数据记录系统研究

霍  炬,杨  明,宋恒杰,王子才   

  1. 哈尔滨工业大学 仿真中心,黑龙江  哈尔滨  150001
  • 收稿日期:2005-01-24 修回日期:2005-03-21 出版日期:2006-03-15 发布日期:2006-03-25
  • 基金资助:
    国家自然科学基金资助项目(60434010)

Research on HLA distributed data collection system based on multi-agent

HUO Ju,YANG Ming,SONG Heng-jie, WANG Zi-cai   

  1. Simulation Cent., Harbin Inst. of Tech., Harbin  150001, China
  • Received:2005-01-24 Revised:2005-03-21 Online:2006-03-15 Published:2006-03-25
  • Supported by:
    Project supported by the National Natural Science Foundation,China(No.60434010).

摘要: 为消除现有数据记录机制对高层体系结构仿真系统性能的不利影响,提出了基于多智能体的高层体系结构分布式数据记录系统的解决方案。该方案采用类似可扩展标记语言的表达形式与Petri网相结合的描述方法,对数据记录过程的行为特征进行形式化描述,实现了对该描述方法的层次语言支持。为满足信息获取的快速高效,以及访问负载的相对均衡,采用分层数据管理与组织机制,实现了对松散耦合的分布数据进行统一管理和更新。最后,针对应用于该系统协调决策过程的强化学习算法进行了探索性研究。

关键词: 高层体系结构, 仿真系统, 分布式数据记录, 数据管理与组织机制, 协调决策

Abstract: To eliminate the adverse effect of existing data collection techniques on the performance of distributed simulation system based on High Level Architecture (HLA), HLA distributed data collection system based on Multi-Agent (MA) was proposed. The system applied a formalization expression method to describe the behavioral characteristics in the process of data collection, and implemented the hierarchy language supports for the descriptions by combining the eXtensible Markup Language(XML) and Petri net. The hierarchy of data management and organization mechanism was realized which could not only effectively support the fast data base access but also ensure access load balancing, and implemented the integrated management and updating of the distributed data. Furthermore, the exploring study of the reinforcement learning algorithm adopted in the MA collaborative decision was offered.

Key words: high level architecture, simulation system, distributed data collection, data management and organization mechanism, cooperative decision

中图分类号: