水库正常蓄水位方案的多目标决策方法研究

发布时间：2018-12-17 19:27

【摘要】：当今世界，水资源短缺问题成为制约全球大多数国家社会经济发展的重要因素。而水利水电工程兴建是合理利用水资源的有效手段。在水库兴建过程中正常蓄水位的确定是一项必要且重要的课题。本文以洪水的资源化为研究背景，通过多目标决策及模糊聚类综合评判方法确定水库正常蓄水位，，主要进行如下工作：一、确定水库正常蓄水位方案优选的指标体系。在系统地分析水库正常蓄水位影响因素的基础上，结合实际工程确定水库正常蓄水位影响评价内容，建立起水库正常蓄水位影响因素评价指标体系。二、对正常蓄水位影响因素指标进行量化及无量纲化处理。本文通过梯形模糊数法进行指标的量化计算，寻找多方面的专家，听取专家对于影响因素的评价，得到客观、真实、可靠的定性指标数据并进行定量化处理。对于定量指标，存在指标量纲不一致、数值没有可比性等因素的影响。本文首先确定每个影响因素的类型，分别定为效益型指标、费用型指标、固定型指标、区间型指标等，通过插值法对定量的指标进行无量纲归一化处理，得到量纲一致的数据。三、运用层次分析——灰色关联法计算综合权重。既体现了专家的主观经验及知识积累对影响因素权重的判断，体现决策者主观喜好；又从客观角度判断权重的大小，体现客观规律对权重的影响。利用主观赋权法与客观赋权法相结合，克服了权重计算的片面性问题，是一种可靠性更高、更加准确的计算权重的方法。四、提出确定水库正常蓄水位的多目标模糊决策方法。本文将利用改进的多目标模糊决策方法对正常蓄水位的优选进行决策，将传统的模糊决策与TOPSIS法结合进行综合决策，得到TOPSIS法评判结果与基本模糊决策结果进行加权平均。提出模糊聚类综合评判法确定水库正常蓄水位。本文把模糊聚类综合评判模型利用在水库正常蓄水位优选中，构造目标差值率矩阵，科学计算隶属度，增加决策结果可靠性，并与多目标模糊决策进行共同验证。五、采用多目标模糊决策方法与模糊聚类综合评判方法对某大型水库进行了水库正常蓄水位方案的优选。通过寻找每个正常蓄水位方案下影响因素的大小，运用梯形模糊数进行指标量化，运用层次分析——关联度法计算综合权重，运用改进的多目标模糊决策与模糊聚类综合评判进行评价，得到了正常蓄水位的最佳方案。所得结果与实际工程中运用的正常蓄水位方案一致。
[Abstract]:In today's world, the shortage of water resources has become an important factor restricting the social and economic development of most countries in the world. The construction of water conservancy and hydropower project is an effective means to make rational use of water resources. It is necessary and important to determine the normal water level in the process of reservoir construction. In this paper, based on the resources of flood, the normal reservoir water level is determined by multi-objective decision making and fuzzy cluster comprehensive evaluation. The main work is as follows: first, the index system of optimal selection of reservoir normal storage level scheme is determined. Based on the systematic analysis of the influencing factors of the normal water level of the reservoir, the evaluation index system of the influence factors of the normal water level of the reservoir is established by combining with the actual project to determine the content of the evaluation of the influence factors of the normal water level of the reservoir. Second, quantitative and dimensionless treatment of the influencing factors of normal water level. In this paper, the quantitative calculation of indicators is carried out by trapezoidal fuzzy number method, and experts from various fields are searched for, and the evaluation of influencing factors by experts is listened to. The objective, true and reliable qualitative index data are obtained and the quantitative processing is carried out. For quantitative indicators, there are some factors such as inconsistent dimensionality and no comparability. In this paper, we first determine the types of each influencing factor, which are defined as benefit index, cost index, fixed index, interval index and so on. The dimensionless normalization of the quantitative index is carried out by interpolation method, and the dimensionally consistent data are obtained. Third, use the analytic hierarchy process-grey correlation method to calculate the comprehensive weight. It not only reflects the subjective experience of the expert and the judgment of the weight of the influencing factors, but also reflects the subjective preference of the decision makers, and judges the weight from the objective angle, and reflects the influence of the objective law on the weight. By combining subjective weight method with objective weighting method, the one-sidedness problem of weight calculation is overcome, and it is a more reliable and accurate method to calculate weight. 4. A multi-objective fuzzy decision-making method for determining the normal reservoir water level is put forward. In this paper, the improved multi-objective fuzzy decision method is used to make decision on the optimal selection of normal water level, and the traditional fuzzy decision making is combined with TOPSIS method for comprehensive decision making, and the weighted average between the evaluation result of TOPSIS method and the basic fuzzy decision result is obtained. A fuzzy cluster comprehensive evaluation method is proposed to determine the normal reservoir water level. In this paper, the fuzzy cluster comprehensive evaluation model is used to construct the target difference rate matrix in the optimal selection of the normal reservoir water level, to scientifically calculate the membership degree, to increase the reliability of the decision result, and to verify it with the multi-objective fuzzy decision making. Fifthly, the multi-objective fuzzy decision method and fuzzy cluster comprehensive evaluation method are used to select the normal water level of a large reservoir. By looking for the size of the influencing factors under each normal water level scheme, the trapezoidal fuzzy number is used to quantify the indexes, and the analytic hierarchy process is used to calculate the comprehensive weight by using the method of correlation degree. The improved multi-objective fuzzy decision making and fuzzy clustering comprehensive evaluation are used to evaluate the optimal scheme of normal water level. The results are consistent with the normal water level scheme used in practical engineering.
【学位授予单位】：河北农业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TV697.1

【参考文献】