【作者】 谭伟；

【导师】 冯仲科；

【作者基本信息】 北京林业大学 ， 森林经理学， 2005， 博士

【摘要】 人工造林一直是基层林场(乡镇林业站)生产经营活动的主体,尤其在当前国内外普遍重视森林生态环境的前提下,大力提倡并开展人工造林活动是森林资源增加与森林质量提高的基本条件,是森林资源补充和生态环境建设的主要途径,是实现林业可持续发展的根本。随着我国林业生态建设的重大举措—六大工程的实施,造林规划显得更加重要。造林决策支持模型是在当前地理信息系统、决策支持系统、专家系统、数据库等技术发展的基础上,结合造林规划设计的具体目标而提出并建立的,是造林规划决策智能化、自动化、科学化、合理化的基础,为造林决策支持系统的建立提供理论依据。 针对基层林场(乡镇)造林规划决策的一般过程—造林作业区选择、立地类型划分与评价、造林树种的选择、造林技术措施提出决策目标,从各决策目标的影响因素与具体特点出发,研究建立了造林作业区决策支持模型、立地决策支持模型、树种决策支持模型、造林技术措施决策支持模型。并以北京市房山区为例,论述了模型建立的基本原理、模型的基本组成及模型运行的模式,研究成果如下: (1)首次系统性地提出并建立造林决策支持模型,其由造林作业区决策支持模型、立地决策支持模型、树种决策支持模型、造林技术措施决策支持模型组成。 (2)创新性地提出运用系统论、集合论的基本思想构建决策支持模型的基本结构,其结构为集合-元素-属性-属性集合。集合由元素组成,元素是集合的基本单元,属性是元素在某一方面的性质,元素的属性组成属性集合。在决策支持模型中,元素体现为决策对象某一方面的性质,是决策对象所处的条件(决策限制条件);属性是决策对象在某一方面决策目标的取值(决策结果)。 (3)提出模型的基本运行模式—映射,所有的决策对象的决策结果都是决策方案对决策对象所处条件的映射。体现在两方面,一为集合内元素-属性的映射,表示定性或定量数据的知识推理,通常表现为多对多、一对多、多对一形式,映射的结果有多个解;一为数学函数映射,表示定量数据的等式运算,体现为多对一(多元)、一对一(一元)关系,其结果为单解。 (4)提出通用型推理引擎建立的基本原理,为基于数据库的SQL语句。在集合论思想的支持下构造的决策支持模型,基本结构是:元素、属性、映射,分别组成决策支持模型的数据库、知识库、模型库。对于任意地区的造林规划问题,数据和知识是不确定的,而数据对知识的映射关系则是确定的,将这种映射关系体现在程序中,以模型体的形式存储下来,则构成通用型推理引擎。 (5)运用嵌入式GIS二次开发工具—MapObjects,在Visual C++的软件开发平台下,集成GIS的空间数据管理与分析等功能、专家系统的专家知识和知识推理法则、DSS于一体,建立造林决策支持模型,是GIS在林业上应用的典型实例,是GIS应用的高级阶段。更多还原

【Abstract】 Afforestation is the principal part of production and management of a forestry centre (forestry station), especially those whose general attention is paid to ecological environment of forest both at home and abroad. Promoting and developing afforestation in a more cost-effective manner is primary for increase of forest resources and improvement of forest quality. It is also the main approach to supplementing of forest resource and constructing ecological environment and realizing forestry’s sustainable development as well. With the carrying out of the six major projects, which is the great action of our country in forestry’s ecological construction, it seems more important to design afforestation operation. Afforestation decision support model (ADSM), based on the development of GIS (Geographic Information System), DSS (decision support system), ES (expert system) and DB (database technique), is proposed and set up with a concrete goal of planning and designing afforestation. It is the foundation of intelligent, automatit, scientific and rationalized process in planning and decision of afforest and can offer the theoretical foundation for the setting up of afforestation decision support system.A decision goal is put forward according to the general course of afforestation planning in a forestry centre: choosing afforestation site, estimating site type, selecting tree species and afforestation techniques. As a result, the decision support model of afforestation site, the model of site type, the model of tree species, and the model of afforest technology have been studied and set up according to the influencing factors and concrete characteristics of every decision item. This paper shows the pattern of models’ operation, the principles and constructure of the models by taking Fangshan District of Beijing as an example. The research results are follows:(1) ADSM was proposed and set up for the first time, consisting of afforestation site decision support model, site type decision support model, afforestation species decision support model and afforestation technique decision support system.(2) Systematic theory and set theory were innovatively suggested for construction of decision support model, of which the structure is aggregate-element-attribute-attribute set. The aggregate is made up of many elements, which are elementary cells of it. The attribute refers to the characters of an element in a certain respect, and when such characters are put together, attribute set is formed. In decision support model, an element shows characters in a certain respect of the decision object. It also shows the conditions of the decision target (the limited conditions of decision), while attribute is the value of the decision goal in a certain respect of target decision (decision result).(3)Mapping - the basic operational mode of model is proposed. All the decision results of the target are the mapping of its conditions determined by the decision scheme. Those are demonstrated in the following two respects: the mapping for the element-attribute within the set, which shows the knowledge reasoning of determined nature or quantitative data, generally appearing in the forms of plurality vs. plurality, single vs. plurality or plurality vs. single with more than one result; and the mapping for the mathematics function, which shows the quantitative equality operation of data, in the form of plurality vs. single (polyhydric), or single vs. single (one) with only one result.(4)The general reasoning engine, consisting of Structured Query Language (SQL) based on set theory and techniques of DB, was put forward. The basic structure of the decision support model: element of set, attribute of element and mapping between them, is related to the database, knowledge base and model house. To the afforestation planning and designing problems in any certain area, the data and knowledge are uncertain, while the mapping between them is certain. This kind of relation is reflected in the procedure and stored in the pattern of model, and therefore the general更多还原