基于多智能体的GIS成矿预测模型研究

项目名称： 基于多智能体的GIS成矿预测模型研究

项目编号： No.41302261

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 天文学、地球科学

项目作者： 周林立

作者单位： 中国科学院合肥物质科学研究院

项目金额： 25万元

中文摘要： 应用新理论、新技术和新方法进行成矿预测从而有效地发现新矿床是解决矿产资源紧缺矛盾的有效途径。近年来，国内外一些研究成果表明：将非线性理论和复杂性理论应用于成矿预测具有良好的发展前景及广泛的应用空间。多智能体系统源于复杂系统适应理论及分布式人工智能，在复杂空间建模中具有无可比拟的优势。本项目拟在研究传统GIS成矿预测模型的基础上，将GIS成矿预测模型这一复杂系统模块化以智能体为单位的为多智能体组织，研究确定各智能体的功能职责、实现策略、工作过程及交互协商机制，并结合SWARM平台与MAPGIS平台进行模型的构建。建立的成矿预测模型将GIS和多智能体有机结合，通过智能体之间的交互协商表达成矿信息与成矿间的复杂关系，具有知识累积、智能化推理与决策功能，能有效提高成矿预测的精度。多智能体系统的引入，是矿产资源评价"智能化"的重要尝试，对于提高GIS矿产资源评价的水平有十分重要的意义。

中文关键词： 多智能体；成矿预测；知识库；空间决策；Jade平台

英文摘要： Application of new theories, new technologies and new methods of mineralization forecast to find new deposits is an effective way to solve the shortage of contradictions in the mineral resources. In recent years, some research results indicate that: nonlinear theory and complexity theory applied to the metallogenic prediction has good prospects and widespread use space. Multi-agent systems stems from the theory adapts to complex systems and distributed artificial intelligence.It has an unparalleled advantage in the modeling of complex spatial problems.Based on the study of traditional GIS metallogenic prediction model,this project first converts the complex system to multi-agent organization, then studies the achieve strategy, work process and interactive consultation mechanism of the agent, finally it builds metallogenic prediction model based on SWARM platform combined with MAPGIS platform. The established metallogenic prediction model which combines GIS and multi-agent and expresses complex relationships between the mineralization information and mineralization through the interaction has the accumulation of knowledge, intelligence, reasoning and decision-making functions. It can effectively improve the accuracy of metallogenic prediction. The application of multi-agent system is an attempt of intelligent

英文关键词： multi-Agent；metallogenic prognosis；knowledge base；spatial decision；jade plateform