【作者】 于苏俊；

【导师】 杨立中；

【作者基本信息】 西南交通大学 ， 市政工程， 2007， 博士

【摘要】 可持续土地利用规划是结合规划区域实际状况,确定出社会、经济和生态诸方面的规划目标并加以协同的过程。为能使可持续土地利用理念能真正融入土地利用规划之中,要求在规划方式和方法上做出相应的改进和创新。可持续土地利用规划摒弃那种类似于部门规划或耕地保护专项规划的做法,强调土地供给要充分满足经济发展的需要,以使土地利用效率最大化。而要顾及这点,需要进行经济发展和土地利用变化关联的分析,揭示经济增长与产业用地及其两者结构的内在联系。论文对基于经济增长的产业用地结构预测研究方法和程序作了有益的探讨。土地利用变化与生态环境演化关联分析是纳入生态目标的关键步骤。论文对土地利用生态安全性研究内容和程序提出了一套思维方式,即以生态区域单元为基础进行特定生态安全性问题诊断,构建某种土地利用变化与某种生态环境影响之间的数量关系模型,并将这种数量关系转化为决策变量和约束条件,纳入土地利用数量结构决策模型之中。寻求符合区域条件约束下最佳或满意的土地利用决策方案是土地利用规划的主要内容。囿于数学手段的限制,传统求解方法并非整体寻优并且属于静态决策。对具有不可公度和相互冲突的多目标协调问题,这种方式不能获得整体最优解。本文提出基于遗传算法的多目标模糊优选动态规划决策分析方法,即利用遗传算法的内在并行机制及其全局优化的特性,解决各时序段多目标优化问题,再由模糊优选对时序上各pareto解集多目标动态优选,最终获得具有现实意义的全局最优决策序列。土地利用规划的不确定性由土地资源未来需求预测的不确定性引起的,后者又是由人口、各种消费需求、经济增长趋势以及新生产业等社会经济因素变化的预测的不确定性引起。论文引入盲数理论对这种不确定性进行处理,由此产生不同置信度下的预测值,预测区间端点值可分别生成两套可调控变量值,用它们构成用地规模变化的弹性调控区间。更多还原

【Abstract】 Sustainable land-use programming can be defined as a process of determining social, economic and ecological multi-objective and managing to coordinate them. It is requested that the programming approaches and methods get improved and innovated for the sake of embodying sustainable land-use concept in the programming.Sustainable land-use programming emphasizes making land satisfy need of whole economic development, instead of those practices similar to plantation safeguard or section project, making land utility efficiency maximize. The correlation analysis between economic development and land-use would be performed to open out the internal logic between economic growth and industry land-use provided this target is asked for. Thus, the forecast method and process for land-use based on economic growth was probed into in this paper.The correlation analysis between ecological evolvement and land-use configuration is a committed step of bringing ecological objective into programming. A set of thinking manner and working process was suggested in this paper. Namely, through building up model between land-use configuration and ecological environment impact, based on ecological subarea in which the issues of ecological security could be diagnosed, the quantitative relation derived from the same is employed as decision variable and constrained condition in structural optimization for land-use model.The main task of land-use programming would be to search optimal or well-pleasing decision scheme of land-use under the conditions of according with regional condition. Limited by traditional math means, whole optimizing from those multi-objective issues characterized with bad commensurability and conflict one another is rather unsuccessful.A decision-analysis method of genetic algorithm + multi-objective fuzzy dynamic optimum was set. Its process can be described as followings:①Solve pareto disaggregate with inherent parallel optimize mechanism of genetic algorithm for each period of time;②Perform multi-objective fuzzy dynamic optimum operation to pareto disaggregate in each period of time, eventually gaining the sequence of global optimize decision-making scheme.The uncertainty of land-use programming is derived from that of future requirement of land amount, and the latter from that of forecast of social and economic growth and changes, including population, consume requirement, economic and rebirth industry. All uncertainty could be effectively dealt with by introducing blind math method into traditional forecast model. Furthermore, since the forecast model based on the blind math can generate multiple predicted values under different confidence level, two set of adjustable variable values corresponding to both extreme point values of prediction interval can easily be gained, composing elasticity adjustable range.更多还原