导　　师： 边馥苓

学科专业： 081602

授予学位： 硕士

作　　者： ;

机构地区： 武汉大学

摘　　要： 地理信息系统在现代社会中的作用越来越重要，这主要是因为地理信息系统不仅可为我们提供丰富的空间信息，而且通过其强大的空间分析能力为决策提供支持信息。地理信息系统为决策支持提供了强大的数据输入、存储、检索、显示的工具，但在分析模拟和推理方面的功能还比较薄弱，本质上是一个数据丰富但知识贫乏的系统，在解决复杂空间决策问题上缺乏智能推理功能。所以，为解决复杂的空间决策问题，需要在地理信息系统的基础上开发智能决策支持系统/(Intelligent Spatial Decision Support Systems，简称ISDSS/)，用于数据获取、输入、存储、分析和输出；用于知识表达与推理；用于自动学习、系统集成、人机交互。它应该能够对结构化或非结构化的知识进行表达与推理。该系统应该以人工智能和知识工程作为知识表达和推理的手段；以软件工程作为系统开发和管理的工具；以空间信息系统作为数据处理和显示的基础框架，并使以上各部分以无缝的方式结合在一起，构成一个完整的空间决策支持系统 对于ISDSS来说，符号方法仍是一种重要的知识表达与推理手段。而对于具有大规模并行分布式结构的知识来说，采用人工神经元网络方法则会比符号方法有较大的优越性。不管ISDSS应用符号方法、神经网络或者混合方法，制约ISDSS发展的瓶颈都将是空间知识的自动获取。从空间数据库中发现/(学习/)知识的能力是衡量一个ISDSS成功与否的重要标志。因此研究人工神经元网络以及其它智能方法在空间知识的获取、表达以及推理中的应用，对于解决上述问题具有重要意义。本文针对人工神经元网络在空间智能决策支持系统中的应用进行了研究，主要工作有如下几方面： 1．人工神经元网络用于空间数学模型的建立 Geographic information system is more and more important in modern society, because the geographic information system not only provides the rich spatial information for us, but also provides the supporting information for us through its formidable spatial analysis ability. The geographic information system has provided the formidable data capture, storage, display's tool for the decision support, but its analysis and inference function is quite weak, essentially it is a data rich but the knowledge deficient system which lacks the intelligent inference function to solve complex spatial decision support problem, therefore, need to develop the Intelligent Spatial Decision Support Systems/(ISDSS/) based on geographic information system, it will be used in the data capture, input, storage, analysis and output; in the knowledge representation and inference; in automatically learning; system integration; man-machine interactive. As for ISDSS, the symbolic approaches were still one kind of important knowledge representation and the inference method. But the artificial neural network which has the massively parallel distributional structure has the superiority which the other means does not have. No matter which approaches ISDSS does apply, the bottleneck which restricts the ISDSS development will be the spatial knowledge acquire, so the ability of data mining and discovery knowledge from spatial database is very important. Therefore studies about the artificial neural network as well as other intelligent methods in the spatial knowledge acquire, representation and inference is vital significance for the problem above. In this article the mainly focus on is the application of artificial neural network in the spatial intelligent decision support systems. The main work has the following aspects: 1. The spatial mathematical models building based on the neural network BP neural network model with hide units may estimate any continuous function by the any precision, so the artificial neural ne

关 键 词： 空间决策支持系统 空间数学模型 空间知识表达与推理 人工神经网络 空间数据挖掘

领　　域： [天文地球] [天文地球]