基于图模型的反事实推理在决策后评估中的应用

中国管理科学 ›› 2012, Vol. 20 ›› Issue (6): 160-166.

基于图模型的反事实推理在决策后评估中的应用

胡笑旋^1,2, 杨善林^1,2, 华娟^1,2

1. 合肥工业大学管理学院,安徽合肥 230009;
2. 合肥工业大学过程优化与智能决策教育部重点实验室,安徽合肥 230009

收稿日期:2011-02-25 修回日期:2012-06-26 出版日期:2012-12-29 发布日期:2012-12-28
基金资助:
国家自然科学基金资助项目(70801024,71001032)

Graph-based Counterfactual Reasoning for Ex-post Decision Evaluation

HU Xiao-xuan^1,2, YANG Shan-lin^1,2, HUA Juan^1,2

1. School of Management, Hefei University of Technology, Hefei 230009, China;
2. Key Laboratory of Process Optimization and Intelligent Decision-making, Hefei University of Technology,Hefei 230009,China

Received:2011-02-25 Revised:2012-06-26 Online:2012-12-29 Published:2012-12-28

摘要/Abstract

摘要： 决策后评估对衡量决策的效果具有重要作用。本文引入反事实推理方法实现对决策结果的对比后评估。首先提出了带有后评估环节的规范性决策流程,然后阐述了使用因果影响图模型进行反事实推理的步骤,定义了自满意度和相对满意度两个指标,通过反事实推理的结果计算指标的分值,从而实现对决策的评价。最后使用例子展示了计算过程,表明了该方法是可行和有效的。

关键词: 影响图, 决策分析, 评估, 反事实推理, 因果模型

Abstract: Ex-post decision evaluation plays an important role in measuring the effects of decisions.In this paper, counterfactual reasoning method is used to implement pairwise decision evaluation.First the normative decision analysis process with post-decision evaluation is proposed, and then the steps of doing counterfactual reasoning with causal influence diagrams are illustrated.Two measures are defined, one is self-satisfaction ratio and the other is relative satisfaction ratio.The values of the measures that acquired through counterfactual reasoning are the results of the assessment.Finally, an example is given to illustrate the evaluation process, which shows the feasibility and validity of the proposed method.

Key words: influence diagram, decision analysis, evaluation, counterfactual reasoning, causal model

中图分类号:

C934

胡笑旋,杨善林,华娟. 基于图模型的反事实推理在决策后评估中的应用[J]. 中国管理科学, 2012, 20(6): 160-166.

HU Xiao-xuan,YANG Shan-lin,HUA Juan. Graph-based Counterfactual Reasoning for Ex-post Decision Evaluation[J]. Chinese Journal of Management Science, 2012, 20(6): 160-166.

参考文献

[1] Jukic T, Bencina J, Vintar M.Multi-attribute evaluation of e-government projects: slovenian approach[J].International Journal of Information Communication Technologies and Human Development,2012, 1(4): 82-92.
[2] You C J, Lee C K M, Chen S L, et al.A real option theoretic fuzzy evaluation model for enterprise resource planning investment[J].Journal of Engineering and Technology Management,2012, 29(1): 47-61.
[3] 胡华安,田益峰,向卓元,等.物资采购综合评价指标体系质量控制模型应用研究[J].中国管理科学,2004, 12(1): 75-80.
[4] Kahneman D, Tversky A.Judgement under uncertainty: Heuristica and biases[M]//The simulation heuristic. New York:Cambridge University Press, 1982.
[5] Balke A, Pearl J.Probabilistic evaluation of counterfactual queries. Proceedings of the 12th national conference on Artificial intelligence.MIT Press,1994.
[6] Heckerman D, Shachter R.A decision-based view of causality//Proceedings of tenth conference on uncertainty in artificial intelligence.San Francisco: 1994.
[7] Heckerman D, Shachter R.Decision-theoretic foundations for causal reasoning[J].Journal of Artificial Intelligence Research,1995,(3): 405-430.
[8] Dawid A P.Influence diagrams for causal modeling and inference[J].International Statistical Review,2002, 70(2): 161-189.
[9] Kjaerulff U B, Madsen A L.Bayesian networks and influence diagrams: a guide to construction and analysis[M].Springer, 2007.
[10] Lin Yonghuang, Lin C C, Tyan Y Y.An integrated quantitative risk anslysis method for major construction accidents using fuzzy concepts and influence diagram[J].Journal of Marine Science and Technology,2011, 19(4): 383-391.
[11] 董志强,赵勇.库存问题的子值结点决策影响图方法[J].中国管理科学,2003, 11(2): 50-54.
[12] 胡笑旋,杨善林,张强.支持群决策过程建模的层次影响图模型[J].系统工程理论与实践,2011, 31(7): 1312-1319.
[13] Pearl J.Causality: models, reasoning and inference[M]. 2nd ed.New York: Cambridge University Press, 2009.
[14] Pearl J.Probabilistic reasoning in intelligent systems: networks of plausible inference[M].San Mateo:Morgan Kaufmann, 1988.
[15] Nadkarni S, Shenoy P P.A causal mapping approach to constructing Bayesian networks[J].Decision Support Systems,2004, 38: 259-281.