中文摘要：

针对黄河流域水资源开发利用具有多目标特征,研究引入决策支持向量机,利用升维和线性化的建模思路,实现对决策者偏好结构的自学习,求解全局最优解、回归多属性决策的隐性效用函数。研究将SVM学习功能与案例推理技术相结合,应用专家经验对决策方案的归类识别以及SVM决策偏好再现的启发模式,实现对决策目标系统的自动建模和推理决策功能,通过定量分析和定性推理的结合,弱化了决策过程中人为的因素,实现了流域水资源利用决策的智能化。通过对黄河流域水资源利用规划水平年案例研究,与专家决策的对比验证了模型和方法智能化和精确性。

英文摘要：

In order to solve the problem of the multi-object problem of water resources exploitation and utilization of Yellow River, we introduce the support vector machine for decision-making and apply the modelling idea of dimension promotion and linearization. We achieve self-study preference structure of the decision-making and seek for the best solutin and regress utility function of the multiple attribute decision making. We combine SVM self-study and reasoning technology, apply the expert＇s experience of classifying scheme and SVM enlighten function of decision making reproduction. We achieve automatic modeling to object of decision making, unite the quantitative analysis and qualitative ratiocination, in course of decision-making allay the influence of artificial factor and obtain the intelligent mode in water resources utilization. Through test and study of water resources development and utilization scheme of 2010 and 2020 year-level of Yellow River, contrast with outcome of the expert decision, the accuracy and intelligent level of the SVM model is validated.