• 关键词：ANN  水资源再生能力  水资源社会再生能力  虚拟城市  黄河流域

• 基金项目：国家自然科学基金(70273005);国家高技术研究发展项目(2003AA601010);国家自然科学基金重点项目(No.50239020)

• 曾维华
• 北京师范大学环境学院 水环境模拟国家重点实验室, 北京 100875

• 杨志峰
• 北京师范大学环境学院 水环境模拟国家重点实验室, 北京 100875

• 祝捷
• 北京师范大学环境学院 水环境模拟国家重点实验室, 北京 100875

• 摘要：为了在城市水资源评价过程中充分考虑城市水资源的社会循环,更好地揭示城市水资源开发利用的潜力,提出了城市水资源社会再生能力概念,并分别建立了城市水资源社会绝对再生能力与相对再生能力的评价指标体系;在此基础上,建立了一种全新的城市水资源评价模型——基于人工神经网络的城市水资源社会再生能力评价模型.该模型在构造评价标准的过程中,通过"虚拟城市"构造模型训练样本;由此将评价问题转化成人工神经网络所擅长的分类问题,并赋予评价结果实际的内涵,使得评价结果更直观且更易于理解.最后,以黄河流域的主要城市为研究对象,对其绝对水资源再生能力与相对水资源再生能力分别进行评价;并与传统的灰色关联方法进行对比;评价结果与实际相吻合.

• Abstract：The urban water resources system is a dynamic system, consisting of climate and water source condition, water supply and drainage works capacity, city property, population, living and technical level, economic strength and industrial structure of city. Thereby, the evaluation of urban water resources includes not only the water resources itself such as water quality and quantity but also the utilization, allocation, supervision and management level; in other words, not only the evaluation of the natural resources but the whole social and economic system as well. The concept of social renewability of urban resources and its assessment indexes was put forward, and ANN-based assessment model of social renewability, a new assessment model of urban water resources, was established in order to consider the social cycle of urban water resources adequately, reveal the exploitation and utilization potential of water resources in the cities, and understand the importance of social renewability of water resources in city water resources assessment. The other characteristic of the model is the construction of the evaluation standard. The maximum and minimum of the assessment indices of the cities in whole country was assembled separately forming two extreme “virtual cities" of highest and lowest grade (5 and 0). One virtual city had the maximum of indices and the other has the minimum. Then the distance between the maximum and minimum was decided, and a range of virtual cities as the training samples of BPnetworks independent upon experience of the appraisers was constructed. In this way, the assessing problem was transformed to a classification problem which ANN was adept in. And the assessment result was more directly perceived through the sense and easy to understand.Furthermore, the model was applied to analyze and assess the main cities in the Yellow River basin, and was compared with the traditional method: Gray-Relation Method. The results of assessment were fit with the practical situation.

• Key words：Artificial Neural Network  water resources renewability  water resources social renewability  Virtual City  Yellow River Basin

• 
  摘要点击次数： 1581 全文下载次数： 2763