不确定环境下基于D数理论的多属性决策研究

发布时间：2018-03-27 23:21

本文选题：多属性决策　切入点：不确定环境　出处：《湖北民族学院》2017年硕士论文

【摘要】：多属性决策在军事、工业工程、经济金融等领域有着广泛应用,一直是学术界研究的热点问题。D数理论,作为D-S证据理论的推广,因为扩展了表达不确定信息条件,近几年广泛应用于多属性决策问题中。然而,随着研究的深入,发现D数理论还存在一些缺陷。因此,本文基于不确定环境下多属性决策问题研究背景,从理论上进一步完善D数理论,并结合变形虫仿生算法将其应用于路径寻优研究中,全文共分为五章。第一章主要介绍D-S证据理论,D数理论与多属性决策的发展历史以及他们在目前国内外研究现状,并对本文所研究的主要内容给出相关准备知识。第二章主要研究目前D数融合规则中不满足交换律的问题,并依据D数自身所包含的已知信息,从评估等级以及相应评估等级的信任程度两个角度分别完善D数融合规则。通过应用于环境评估的实例并将结果与前人文献的对比研究表明,本文提出的方法能有效地解决多个D数融合不满足交换律的问题,并可应用于多属性决策问题。第三章主要研究了D数集成公式中存在的一个不足,因为直接忽略最终评估结果的不完备信息而导致融合信息精确度降低并可能导致决策失误。本章依据已经获得的评估结果并提出改进方法,完善了D数理论的集成公式。通过对摩托车表现评估的实际例子验证表明,完善后的理论提高了决策的辨识度。第四章主要研究了D数理论在路径寻优中的应用。将D数理论应用于路径寻优问题中的道路评估指标冲突、道路属性值不完备以及限制必经道路等情况下,结合仿生算法(变形虫算法),分别建立几种情况下的数学模型,并通过数值实验证明建立的模型能有效搜索到不确定环境下的最优路径,而且算法具有较低的复杂度。第五章总结,对于本文研究工作进行了讨论并对未来研究提出了展望。
[Abstract]:Multi-attribute decision making, which is widely used in military, industrial engineering, economic and financial fields, has always been a hot topic in academic circles. As a generalization of D-S evidence theory, it extends the expression of uncertain information conditions. In recent years, it has been widely used in multi-attribute decision making. However, with the development of research, it is found that there are still some defects in the theory of D number. Therefore, this paper is based on the background of multi-attribute decision making in uncertain environment. The D number theory is further improved theoretically and applied to the path optimization research with the amoeba bionic algorithm. The first chapter mainly introduces the development history of D-S evidence theory and multi-attribute decision making, as well as their current research situation at home and abroad. In the second chapter, we mainly study the problem of not satisfying the commutative law in the current D-number fusion rules, and according to the known information contained in D-number itself, The D-number fusion rules are improved from the evaluation level and the degree of trust of the corresponding evaluation level, and the results are compared with the previous literatures by the example of environmental assessment. The method proposed in this paper can effectively solve the problem that multiple D-number fusion does not satisfy the commutative law, and can be applied to the multi-attribute decision making problem. Because the incomplete information of the final evaluation result is ignored directly, the accuracy of the fusion information is reduced and the decision error may be caused. The integration formula of D number theory is perfected. The actual example of motorcycle performance evaluation shows that, In chapter 4, the application of D number theory in path optimization is studied. In the case of incomplete road attribute value and restricted path, combined with bionic algorithm (amoeba algorithm), the mathematical models of several cases are established respectively. Numerical experiments show that the proposed model can effectively search the optimal path in uncertain environment, and the algorithm has lower complexity. Chapter five summarizes the research work in this paper and puts forward the prospect of future research.
【学位授予单位】：湖北民族学院
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：O225

【参考文献】