KD-ACP计算实验平台的应用模式设计  

Application Pattern Designs of KD-ACP Computational Experiment Platform

作　　者：邱晓刚[1,2] 李祯[1] 张鹏[1] 陈彬[1] 郭刚[1] 

机构地区：[1]国防科技大学信息系统与管理学院,湖南长沙410073 [2]国防科技大学军事计算实验与平行系统研究中心,湖南长沙4100731

出　　处：《指挥与控制学报》2015年第2期164-169,共6页Journal of Command and Control

基　　金：国家自然科学基金(91024030;71303252)资助

摘　　要：针对大规模人工社会计算实验的需求,基于ACP的思想,设计并实现了一种高性能、可扩展、可定制、开放交互的大规模Agent计算实验平台(KD-ACP).从用户使用的角度,该平台提供5个功能模块:人工社会、计算实验、平行控制、计算环境管理和仿真资源管理.在概要阐述了平台框架的基础上,以突发事件应急管理的计算实验为背景,重点讨论了KD-ACP平台的应用模式:通过资源的组合和配置生成能够支持对不同类型的突发事件计算实验的应用系统;通过计算实验的应用来扩展新的资源,并集成进平台来逐步提升平台的能力.目前,KD-ACP平台已用于"甲型H1N1流行事件"、乌鲁木齐"7.5事件"、雾霾舆情演化、生化事件疏散、埃博拉疫情传播等的研究.通过这些应用,平台集成的资源可以初步支持舆情和疫情两类人工社会计算实验.Aiming at the demand of large scale artificial society computational experiment, based on the ACP theory, the paper designs and realizes a large scale Agent based computational experiment platform KD-ACP which is characterized with high performance, extendable, customizable and open interactive. From user＇s point of view, the platform includes 5 function modules： artificial society, computational experiment, parallel control, computational environment management and repository management. Based on the brief illustration of the platform framework, on the background of emergency management computational experiment, the paper focuses on the argument of the application pattern design of KD- ACP platform： through assembling and configuration, generating the application system which supports computational experiment on different kinds of emergency; extending new resources by applying computational experiment, and integrate them into platform to raise the ability of platform gradually. KD-ACP platform has implied in the research of ＂A-H1N1 influenza event＂, Urumchi ＂7.5＂ Riot, public opinion evolution on haze, biochemical event evacuation, and Ebola outbreaks. Through these applications, the platform can preliminarily support public opinion and epidemic computational experiment.

关 键 词：计算实验 人工社会 复杂系统 突发事件管理 系统仿真 

分 类 号：TP311.52[自动化与计算机技术—计算机软件与理论]