AHP在群决策中的应用研究

【作者】 辛杨；

【导师】 荣莉莉；

【作者基本信息】 大连理工大学 ， 管理科学与工程， 2001， 硕士

【摘要】 决策的成功离不开科学的决策方法。对于复杂的决策问题，为了避免失误，提高决策的准确性，需要依靠群组专家的智慧进行群决策，提高决策的质量。本文针对多属性决策问题的决策分析方法一层次分析法（AHP）展开了研究。 论文首先系统地分析了群决策研究的现状和利用AHP进行群决策的优点以及存在的问题，对AHP的基本原理进行了论述，探讨了如何根据实际问题和已有的资料，选择适当的AHP应用模型进行决策分析。 其次，论文对群组AHP中评价结果的合成方法进行了研究，指出了算术平均法和加权平均法的局限，提出专家权重的获得应该从评价专家的角度去考虑。专家对问题的评价要受到专家的知识结构、经验、研究领域等因素的影响，因此如果能够选择合理的评价指标对专家进行评价，可以获得专家权重，以克服其它方法的局限性。本文将通过该方法得到的专家权重定义为专家支持度。利用专家支持度，有效集结每个专家的决策结果，可以提高群决策结果的可靠性和合理性。利用三标度法和最优传递矩阵，解决了AHP中人的判断的模糊性和不确定性问题，以及判断矩阵的一致性调整问题。利用上述三种方法对群组AHP进行了改进。 论文最后论述了模糊层次分析法（FAHP）的思想，重点分析了一种实用的FAHP中专家权重确定方法存在的问题，探讨了如何利用模糊数学的方法处理专家信息，获得FAHP中专家支持度的模糊表示。通过基于改进的群组AHP的应用示例—学术论文网上评价，进行了仿真研究。 AHP作为一种简单且实用决策分析方法，如果能够很好地解决其对人的主观判断的倚重，进行决策时客观、合理地处理专家评价结果，将会提高AHP应用的有效性、科学性。更多还原

【Abstract】 The successful decision-making can’t be obtained without the scientific decision-making methods. For some complicated decision problems, in order to avoid the mistakes and to improve the accuracy, it is needed to rely on the wisdom of a group of experts to make a decision. The thesis studies the analytic hierarchy process (AHP), which is a method for solving decision problems with multiple criteria. The thesis first analyzes the present situation of the researches on group decision-making, and the advantages and disadvantages of using AHP to make group decision are also discussed. The basic theories of ALJP are reviewed and it is discussed how to choose the appropriate application models of AHP to make decision based on the problem at hand and existent data. Then, the thesis studies the methods for composing evaluation results in group-AIIP and points out the limitations of the methods such as arithmetical average method and weighted average method. The acquisition of the experts weights should be considered from an angle of evaluating the experts themselves, because the factors such as the knowledge structure, experience and research fields of experts-could influence the expert’s evaluation to problems. The experts weights can be obtained by choosing the appropriate indexes to evaluate the experts. This method can overcome the limitations of prior methods. The thesis defines the experts-weights obtained by this method as the expert supporting degree. The reliability and reasonability of group decision results can be improved by using the expert supporting degree to compose each expert’s decision result effectively. Using the three-scales-method and the optimal transitive matrix method to solve the problem of fuzzy and uncertain properties of judgement and that of the adjustment of continuity of judgement matrix. Group- AHP is improved by using the stated three methods. Finally, the thesis discusses the theories of FAI-P and mainly analyzes the shortcomings of acquisition method of experts weights in a kind of practical FAHP. It proposes how to dispose the information of experts by fuzzy mathematical methods and obtains the fuzzy expert supporting degree. A simulation study is presented according to an applied example of the eb-based Evaluation of Academic Theses based on the improved group-AHP. As a simple and useful decision-making method, the effectivity and scientificity of AHP will be improved if the subjective judgement of people can be resolved well and the experts evaluation results can be disposed objectively and reasonably.更多还原