在继承传统层次分析/网络分析方法的基础上提出一种新的决策方法——基于有限状态齐次马尔科夫链的网络决策方法MC-ANP. 该方法使用有向图定义决策准则及准则之间的支配关系,通过对准则的两两比较量化支配关系,用马氏链的状态转移描述支配关系. 该方法突出了准则支配关系的合成过程,提供了两种合成模型(其中积合成模型可以彻底解决传统层次分析中出现的逆序问题), 指出传统网络分析中无条件使用Cesaro平均极限求解存在的问题.新方法将决策问题分为两类:有方案的决策问题,解是对方案的排序, 求解过程就是求马氏链状态转移概率方阵的属于特征根1的特定的左特征向量;无方案决策问题,解是对准则的排序,解是马氏链状态转移概率方阵的属于特征根1的右特征向量.
Abstract
On the basis of the traditional AHP/ANP, a new decision making approach — MC-ANP (analytic Network Process based on finite state homogeneous Markov chain) — was proposed. In this approach we have used a direct graph to define the criteria and the dominance relations among them, have quantified the relations of dominations through comparison of criteria, and have~described criteria dominance relations~using the matrix of state transition of the Markov chain. This approach has focused on the synthesis of the dominance relations, has provided two synthetic models (among which the multiply model can be used to solve rank reverse problem throughly), and has pointed out the confusion existed in the traditional ANP using Cesaro mean limit. In the MC-ANP, the decision problems are classified in two categories: i) ranking the alternatives, the process of the solution is to find the specific left eigenvectors of eigenvalue 1 of matrix of state transition probability of the Markov chain; ii) ranking the criteria (no alternative), the solution is the right eigenvector of eigenvalue 1 of matrix of state transition probability of the Markov chain.
关键词
决策方法 /
层次分析(AHP) /
网络分析(ANP) /
Markov链 /
邻接方阵
{{custom_keyword}} /
Key words
decision making /
AHP (analytic hierarchy process) /
ANP (analytic network process) /
Markov chains /
adjacency matrix
{{custom_keyword}} /
中图分类号:
C934
{{custom_clc.code}}
({{custom_clc.text}})
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 赵焕臣,许树柏,和金生. 层次分析法——一种简易的新决策方法[M]. 北京: 科学出版社,1986. Zhao H C, Xu S B, He J S. The Method of AHP — A Simple New Decision Method[M]. Beijing: Science Press, 1986.
[2] 王莲芬,许树柏. 层次分析法引论[M]. 北京: 中国人民大学出版社, 1990. Wang L F, Xu S B. The Introduction to the Analytic Hierarchy Process[M]. Beijing: China Renmin University Press, 1990.
[3] 许树柏. 层次分析法原理[M]. 天津: 天津大学出版社, 1988. Xu S B. The Principle of the Analytic Hierarchy Process[M]. Tianjin: Tianjin University Press, 1988.
[4] Saaty T L. The Analytic Network Process[M]. 2nd ed. RWS Publications, 2001.
[5] Saaty T L. Relative measurement and its generalization in decision making why pairwise comparisons are central in mathematics for the measurement of intangible factors the analytic hierarchy/network process[J]. Rev R Acad Cien Serie A Mat, 2008, 102(2): 251-318.
[6] 许树柏,杨嘉. 浅议ANP[C]// 中国系统工程学会决策科学专业委员会第四届学术年会论文, 2001. Xu S B, Yang J. Introduction to ANP [C]//Paper on the 4th Symposium on the Decision Science Committee of SESC, 2001.
[7] 王莲芬. 网络分析法(ANP)的理论与算法[J]. 系统工程理论与实践, 2001, 21(3): 44-50. Wang L F. The theory and algorithm of analytic network process[J]. Systems Engineering — Theory & Practice, 2001, 21(3): 44-50.
[8] Saaty T L. The seven pillars of analytic hierarchy process[C]//Proceedings of the 5th International Symposium on Analytic Hierarchy Process, Kobe, 1999: 20-23.
[9] 王莲芬,朱枫涛. AHP中逆序问题的综述[J].系统工程理论与实践,1994, 14(6): 1-6. Wang L F, Zhu F T. A summary of rank reversal in AHP[J]. Systems Engineering — Theory & Practice, 1994, 14(6): 1-6.
[10] 刘奇志.层次分析积因子方法的特性及理论基础[C]//中国系统工程学会决策科学专业委员会第四届学术年会论文选集. 决策科学的理论与方法,北京: 海洋出版社, 2001: 19-33. Liu Q Z. The properties and theory of product AHP method[C]// Proceedings of the 4th Symposium on the Decision Science Committee of SESC, Scientific Theories and Methods of Decision-making, Beijing: Ocean Press, 2001: 19-33.
[11] Liu Q Z. Product AHP and it's properties[C]// Proceedings of the 6th International Symposium on the Analytic Hierarchy Process, Berne, 2001: 341-348.
[12] Liu Q Z, Wang Q. Product method of analytic hierarchy process[J]. Asia-Pacific Journal of Operational Research, 1991(8): 135-145.
[13] Liu Q Z, Wang Q. Product method of AHP[C]//Proceedings of International Symposium on the Analytic Hierarchy Process, Tianjin, 1988: 225-231.
[14] 刘奇志.层次分析积因子方法的保序性[J].系统工程学报, 1995(10): 61-70. Liu Q Z. Rank preservation property of product AHP[J]. Journal of Systems Engineering, 1995(10): 61-70.
[15] 刘奇志. 积合成模型是唯一保序合成模型的条件[J]. 数学的实践与认识, 2011, 41(12): 54-60. Liu Q Z. The conditions for multiplicative composition model to be the only rank preservation composition model[J]. Mathematics in Practice and Theory, 2011, 41(12): 54-60.
[16] 刘奇志.网络决策分析的积因子方法[J]. 系统工程的理论与实践, 2004, 24(9): 90-97. Liu Q Z. The product method of analytic network process[J]. Systems Engineering — Theory & Practice, 2004, 24(9): 90-97.
[17] 刘奇志. 几种多属性决策方法的使用分析[C]//中国系统工程学会决策科学专业委员会第七届学术年会论文选集. 决策科学理论与创新, 北京: 海洋出版社, 2007: 142-147. Liu Q Z. Analysis for some multi-attribute decision making methods[C]// Proceedings of the 7th Symposium on the Decision Science Committee of SESC. Scientific Theory and Creation for Decision-making, Beijing: Ocean Press, 2007: 142-147.
[18] 徐光辉. 运筹学基础手册[M]. 北京: 科学出版社, 1999. Xu G H. Handbook of Operational Research Base[M]. Beijing: Science Press, 1999.
[19] 蒋正新,等. 矩阵理论及其应用[M]. 北京: 北京航空学院出版社, 1988. Jiang Z X, et al. Matrix Theory and Application[M]. Beijing: Beijing Aviation Institute Press, 1988.
[20] 曹志浩,等. 矩阵计算和方程求根[M]. 北京: 高等教育出版社, 1983. Cao Z H, et al. Matrix Calculation and Roots of Equations[M]. Beijing: Higher Education Press, 1983.
[21] 宣家骥. 多目标决策[M]. 长沙: 湖南科学技术出版社, 1989. Xuan J J. Multi-objective Decision Making[M]. Changsha: Hunan Science and Technology Press, 1989.
[22] 岳超源. 决策理论与方法[M]. 北京: 科学出版社, 2003. Yue C Y. The Theory and Method of Decision Making[M]. Beijing: Science Press, 2003.
{{custom_fnGroup.title_cn}}
脚注
{{custom_fn.content}}