一种基于多Agent和GIS的麦蚜种群动态模拟方法研究初报  

Development of large-scale spatio-temporal dynamic model of wheat aphid population based on the multi-agent system and geographic information system

作　　者：陈林[1] 程登发[1] 陆庆光[2] 田喆[1] 姜玉英[3] 孙京瑞[1] 张云慧[1] 乔红波[1] 

机构地区：[1]中国农业科学院植物保护研究所 [2]中国农业科学院研究生院,北京100081 [3]全国农业技术推广服务中心,北京100026

出　　处：《植物保护》2006年第6期33-38,共6页Plant Protection

基　　金：国家"973"计划项目(2006CB102007;2002CB111400)

摘　　要：利用Agent模型对昆虫及其寄主、环境之间的关系进行计算机模拟,比传统种群模型更为直观的体现三者之间的信息交互过程。本文建立了麦长管蚜(Sitobionavenae)、小麦(TriticumaestivumL.)、天敌组合、环境4种类型Agent对象,观察了多种组合下的对象行为。以北京地区的历史日值均温、湿度、降雨量数据为主要环境数据,进行了区域模拟验证,与北京地区实际调查数据进行了对比研究。在GIS平台上,以小麦生长季节的气象数据对宁夏、陇东、陕西3个地区进行了模拟,以环境数据进行驱动的4类Agent对象运行,得到了初步模拟结果。With the agent-based theory the computer-simulation of the relationships among insects, hosts and the surroundings showed that the information exchanging process among them was more direct and obvious than that by traditional models or methods. Including aphid population agent, wheat agent, natural enemy unit agent and surrounding agent, a multi-agent system （MAS） was developed in this study and the behaviors of the agents were observed. Upon that it was applied the climatic data of Beijing as input data into the system, such as mean temperature, moisture, and rainfall of one day, in order to simulate the dynamics of every component and to prove the func tions of our model. Thereafter the simulation results were compared with the real data obtained from wheat fields. Furthermore based on the geographic information system （GIS）, the multi-agent system was driven by the Surround ing data in the areas of Ningxia, Shannxi and east of Gansu Provinces. As a result a preliminary outcome was calculated by the multi-agent GIS system. It demonstrated that the agent-based system was effectual in large-scale analysis of pest populations.

关 键 词：MULTI—AGENT System(MAS)多智能体系统 大区预测预报 GIS Logistic曲线 计算机仿真 

分 类 号：S431.9[农业科学—农业昆虫与害虫防治]