基于信息融合的转基因食品安全评估
摘要
转基因食品自问世以来,全球约有2亿多人食用过数千种转基因食品。转基因技术作为一项新兴的技术,其食品的食用安全问题受到各界人士的广泛关注。有试验表明,部分动物食用转基因食品后,会出现各类异常情况甚至是死亡,转基因食品是否对人体有害成为学术界讨论的热点问题。目前,关于转基因食品安全的研究很多,这些研究多是以某一角度作为切入点来评估转基因食品的食用安全,尚未针对转基因食品的多因素特点建立基于信息融合的综合分析平台。因此,建立一个有效融合多属性信息的智能评估平台对我国转基因食品的发展具有重要的现实意义和理论价值。
转基因食品安全评估系统是一个复杂的系统,涉及诸多因素,既有能统计到具体数值的定量数据,又有模糊语言组成的定性指标,各个因素对整体安全评估系统的影响程度有大小之分。本文针对转基因食品的安全评估过程中的诸多问题进行了探讨,引入等级全息建模方法(HHM)、层次分析法(AHP)与证据理论相结合的方法建立多属性信息融合决策模型,并开发了针对转基因食品食用安全的智能评估平台,具体研究内容如下:
(1)建立转基因食品安全评估模型。转基因食品的安全问题涉及到很多方面的影响因素,本文将这些影响因素称为影响指标,各影响指标的选取是通过大量的文献调研和与转基因食品安全研究的专家协商而确定的。文中将影响指标分为几大类一级指标,再将一级指标细分为多个二级指标,部分二级指标包含有三级指标。这些指标包括以统计数据出现的定量指标和以专家的模糊语言判断为依据的定性指标。不同层级的指标构成了转基因食品安全评估模型的框架。
(2)运用AHP方法计算每一层级指标的权重。AHP方法是一种根据相对重要性的比例标度计算出每个影响指标权重的方法。将转基因食品安全评估模型拆分为多个递阶层次结构,请专家给出每一递阶层次结构指标的相对重要程度,得到判断矩阵,求出判断矩阵的特征向量、权重向量和一致性指标CR值。当CR值满足一致性判断条件时即可得出给递阶层次结构的各个指标的权重值。
(3)研究将定性指标和定量指标转化为基本概率指派函数(BPA)的方法。对定性指标,请相关领域专家给出指标的安全等级并给出确信度,从而得到BPA函数;而对于定量指标,利用已有的统计数据样本,将数据划分为多个区域,每个区域分别对应转基因食品的某个安全等级。对某个具体数据,将其与划分区域的分界点进行比较,得到该数据所在区域,从而得到其安全等级,并给出安全等级的BPA函数值。
(4)在承认证据有冲突的基础上,比较前人对冲突证据处理方法的优劣性,本文采用邓勇提出的加权融合算法对冲突证据进行融合,得到最终BPA函数。BPA函数值最大者对应焦元即为安全等级,BPA函数的最大值即为确信度。
(5)使用Matlab设计评估界面,并编程实现转基因食品安全评估系统的软件开发。该系统可实现对不同年份不同地区的各类转基因作物食用安全的综合评估,也可实现对转基因评估系统的各级指标的安全等级评估。
Since the advent of genetically modified food(GM food), global consumption of about200 million people eat several thousand kinds of GM food. As an emerging technology, GMfood safety issues are widespread concerned. Experiments show that some animals eatingGM food will perform abnormal or even die. It becomes a hot academic discussion topic ofwhether GM food is harmful to human.
So far, there are a lot of researches about GM food safety, many of which are based ona point of view as a starting point to evaluate GM food safety ,but all of these researchesfailed to combine with the multi-factor features in GM food system to build models based oninformation fusion analyzing platform. Therefore, the establishment of an effective multi-attribute information integration platform for intelligent assessment of China’s developmentof GM food has important practical and theoretical value.
GM food safety assessment system is a complex system, which involvs many factors,including quantitative statistical data and qualitative indicators composed by fuzzy linguistic.Each factor impacts differently to the overall safety evaluation system of the GM food.This paper aims to establish a multi-property information fusion decision-making modeland then develop an intelligent GM food safety assessment system which is based on theHierarchy Holographic Modeling(HHM) method, Analytic Hierarchy Process(AHP) methodcombined with D-S evidence theory. Specific researches are as follows:
(1) Establish GM food safety assessment model. GM food safety is related to manyfactors which are called as indicators in this paper. The impact indicators of those factors areobtained by searching related documents and consulting specific experts.In this paper, the impact indicators are divided into several first level indicators. Eachfirst level indicator is then broken down into multiple second indicators, some secondaryindicators contain third indicators. These indicators include quantitative statistical data andqualitative indicators by expert’s fuzzy language judgments. Different levels of indicatorsconstitutes the framework of GM food safety assessment model.
(2) Get the indicators’s weights in each level by AHP. AHP method is based on the ratioof relative importance to calculate each impact index weight. The safety assessment of GMfood in the target model is split into several multiple hierarchical structures. Experts will beinvited to give indicator’s relative importance in each hierarchical structure. Then we willget matrix, eigenvector, weight vector and consistency indicator CR value by calculatingthe matrix. The weights of individual indicators will be got when CR value satisfy theconsistency conditions.
(3) Research method how qualitative and quantitative indicators are changed into basicprobability assignment(BPA) of evidence theory. To qualitative indicators, asking relevantexperts give the safety level and their confident to get BPA functions. To quantitative indi-cators, the existing statistical data sample is divided into multiple regions by several cut-offpoints, each region correspond to a safety level of GM food. To a specific data, compared tothe cut-off points, its region will be got resulting in its safety level and BPA function.
(4) On the foundation that evidences have con?icts, the methods have advantages anddisadvantages which are used to deal with con?ict evidence. In this paper, Deng Yong’smethod is used to deal with con?ict evidence to get the final BPA function. The biggest BPAfunction is corresponding to GM food’s safety.
(5) In this paper, we develop an intelligent system to assess GM food safety, based onMatlab programming and graphic user interface(GUI) environment. This system can realizedthe integrated assessment to China’s different areas, different years and different types of GMfood as well as its single indicators respectively.
转基因食品安全评估系统是一个复杂的系统,涉及诸多因素,既有能统计到具体数值的定量数据,又有模糊语言组成的定性指标,各个因素对整体安全评估系统的影响程度有大小之分。本文针对转基因食品的安全评估过程中的诸多问题进行了探讨,引入等级全息建模方法(HHM)、层次分析法(AHP)与证据理论相结合的方法建立多属性信息融合决策模型,并开发了针对转基因食品食用安全的智能评估平台,具体研究内容如下:
(1)建立转基因食品安全评估模型。转基因食品的安全问题涉及到很多方面的影响因素,本文将这些影响因素称为影响指标,各影响指标的选取是通过大量的文献调研和与转基因食品安全研究的专家协商而确定的。文中将影响指标分为几大类一级指标,再将一级指标细分为多个二级指标,部分二级指标包含有三级指标。这些指标包括以统计数据出现的定量指标和以专家的模糊语言判断为依据的定性指标。不同层级的指标构成了转基因食品安全评估模型的框架。
(2)运用AHP方法计算每一层级指标的权重。AHP方法是一种根据相对重要性的比例标度计算出每个影响指标权重的方法。将转基因食品安全评估模型拆分为多个递阶层次结构,请专家给出每一递阶层次结构指标的相对重要程度,得到判断矩阵,求出判断矩阵的特征向量、权重向量和一致性指标CR值。当CR值满足一致性判断条件时即可得出给递阶层次结构的各个指标的权重值。
(3)研究将定性指标和定量指标转化为基本概率指派函数(BPA)的方法。对定性指标,请相关领域专家给出指标的安全等级并给出确信度,从而得到BPA函数;而对于定量指标,利用已有的统计数据样本,将数据划分为多个区域,每个区域分别对应转基因食品的某个安全等级。对某个具体数据,将其与划分区域的分界点进行比较,得到该数据所在区域,从而得到其安全等级,并给出安全等级的BPA函数值。
(4)在承认证据有冲突的基础上,比较前人对冲突证据处理方法的优劣性,本文采用邓勇提出的加权融合算法对冲突证据进行融合,得到最终BPA函数。BPA函数值最大者对应焦元即为安全等级,BPA函数的最大值即为确信度。
(5)使用Matlab设计评估界面,并编程实现转基因食品安全评估系统的软件开发。该系统可实现对不同年份不同地区的各类转基因作物食用安全的综合评估,也可实现对转基因评估系统的各级指标的安全等级评估。
Since the advent of genetically modified food(GM food), global consumption of about200 million people eat several thousand kinds of GM food. As an emerging technology, GMfood safety issues are widespread concerned. Experiments show that some animals eatingGM food will perform abnormal or even die. It becomes a hot academic discussion topic ofwhether GM food is harmful to human.
So far, there are a lot of researches about GM food safety, many of which are based ona point of view as a starting point to evaluate GM food safety ,but all of these researchesfailed to combine with the multi-factor features in GM food system to build models based oninformation fusion analyzing platform. Therefore, the establishment of an effective multi-attribute information integration platform for intelligent assessment of China’s developmentof GM food has important practical and theoretical value.
GM food safety assessment system is a complex system, which involvs many factors,including quantitative statistical data and qualitative indicators composed by fuzzy linguistic.Each factor impacts differently to the overall safety evaluation system of the GM food.This paper aims to establish a multi-property information fusion decision-making modeland then develop an intelligent GM food safety assessment system which is based on theHierarchy Holographic Modeling(HHM) method, Analytic Hierarchy Process(AHP) methodcombined with D-S evidence theory. Specific researches are as follows:
(1) Establish GM food safety assessment model. GM food safety is related to manyfactors which are called as indicators in this paper. The impact indicators of those factors areobtained by searching related documents and consulting specific experts.In this paper, the impact indicators are divided into several first level indicators. Eachfirst level indicator is then broken down into multiple second indicators, some secondaryindicators contain third indicators. These indicators include quantitative statistical data andqualitative indicators by expert’s fuzzy language judgments. Different levels of indicatorsconstitutes the framework of GM food safety assessment model.
(2) Get the indicators’s weights in each level by AHP. AHP method is based on the ratioof relative importance to calculate each impact index weight. The safety assessment of GMfood in the target model is split into several multiple hierarchical structures. Experts will beinvited to give indicator’s relative importance in each hierarchical structure. Then we willget matrix, eigenvector, weight vector and consistency indicator CR value by calculatingthe matrix. The weights of individual indicators will be got when CR value satisfy theconsistency conditions.
(3) Research method how qualitative and quantitative indicators are changed into basicprobability assignment(BPA) of evidence theory. To qualitative indicators, asking relevantexperts give the safety level and their confident to get BPA functions. To quantitative indi-cators, the existing statistical data sample is divided into multiple regions by several cut-offpoints, each region correspond to a safety level of GM food. To a specific data, compared tothe cut-off points, its region will be got resulting in its safety level and BPA function.
(4) On the foundation that evidences have con?icts, the methods have advantages anddisadvantages which are used to deal with con?ict evidence. In this paper, Deng Yong’smethod is used to deal with con?ict evidence to get the final BPA function. The biggest BPAfunction is corresponding to GM food’s safety.
(5) In this paper, we develop an intelligent system to assess GM food safety, based onMatlab programming and graphic user interface(GUI) environment. This system can realizedthe integrated assessment to China’s different areas, different years and different types of GMfood as well as its single indicators respectively.
引文
[1] D.L. Hall, Mathematical Techniques in Multisensor Data Fusion, ArtechHouse, 1992.
[2] D.L. Hall, J. Llinas, An introduction to Multisensor Data Fusion, Proc of the IEEE, pp.6-23, 1997.
[3] E. L. Waltz, Data Fusion for C3I, PaloAlto,CA:EWCommunications, pp.217-226, 1986.
[4] W.S. Gesing, D.B. Reid, An integrated multisensor aircraft track recovery system for remote sensing,IEEETrans.Automat.Contc, Vol. 28, No. 3, pp.356-363, 1983.
[5] S.A. Shafer, A. Stentz, C.E. Thorpe, Architecture for sensory fusion in amobile robot, Proc IEEEInt. Conf. Robotics and Automat, SanFrancisco, pp.2002-2011, 1986.
[6] D.J. Kriegman, l.E. Triend, T.O. Binfork,Stereo vision and navigation in buildings for mobile robots,IEEETrans.Robot.Automat, Vol. 5, No. 6, pp.792-803, 1989.
[7] K. Demirbas,Integration of tactile sensors and machine vision for control of robotic manipulators,Robots 9 Conf. Proc, pp. 37-71, 1987.
[8] Konig, A ,et al,Assessment of the safety of foods derived from genetically modified (GM) crops, Foodand chemical toxicology, Vol. 42, NO. 7, pp.1047-1088, 2004.
[9] Engel KH, Frenzel T, Miller A,Current and future benefits from the use of GM technology in foodproduction,toxicology letters,Vol. 127, NO. 1-3, pp.329-336, Feb. 2002.
[10] Zdunczyk Z,In vivo experiments on the safety evaluation of GM components of feeds andfoods,Journai of animal and feed sciencies,Vol. 10, NO. Supplement 1, pp.195-210, 2001.
[11] Beachy Roger N,Current status and future prospects of virus-resistant GM plants in strengtheningfood security,Indian journal of virology, Vol. 19, NO. 1, pp.51-51, Jan. 2008.
[12] La Mura Maurizio, Allnutt Theodore R., Greenland Andrew, et al, Application of QUIZ for GMquantification in food, Food Chemistry, Vol. 125, NO. 4, pp.1340-1344, Apr. 2011.
[13] Nielsen Annika Porsborg, Lassen Jesper, Sandoe Peter, Public participation: democratic ideal orpragmatic tool? The cases of GM foods and functional foods, Public understanding of science, Vol.20, NO. 2, pp.163-178, Mar. 2011.
[14] Cook AJ, Kerr GN, Moore K, Attitudes and intentions towards purchasing GM food, Journal ofeconomic psychology, Vol. 23, No. 5, pp.557-572, Oct. 2002.
[15] Gaskell G, Allum N, Wagner W; et al., GM foods and the misperception of risk perception, Riskanalysis, Vol. 24, No. 1, pp.185-194, Feb. 2004.
[16]金敬强,武富春,信息融合技术的发展现状与展望,电脑与应用, 2006: 50-52.
[17] Varshney PK, Multisensor data fusion, Electronic and communication engineering journal, Vol. 9,NO. 6, pp.245-253, Dec. 1997.
[18] Benaskeur Abder Rezak, Rheaume Francois, Adaptive data fusion and sensor management for mil-itary applications, Aerospace science and technology, Vol. 11, NO. 4, pp.327-338, May. 2007.
[19] Putvinskaya N, Bulasheva N, Cole G, et al., Information technology systems for fusion industry anditer project, Fosion science and technology, Vol. 47, NO. 3, pp.806-813, Apr. 2005.
[20] Lou KN, Lin CJ, An intelligent sensor fusion system for tool monitoring on a machining centre,International journal of advanced manufacturing technology, Vol. 13, NO. 8, pp.556-565, 1997.
[21] Fang YJ, Masaki I, Horn B, Depth-based target segmentation for intelligent vehicles: Fusion ofradar and binocular stereo, IEEE Transactions on intelligent transportation systems, Vol. 3, NO. 3,pp.196-202, Sep. 2002.
[22] Waxman AM, Fay DA, Rhodes BJ, et al., Information fusion for image analysis: Geospatial founda-tions for higher-level fusion , Proceeding of the fifth international conference on information fusion,Vol. 1, pp.562-569, 2002.
[23]李德毅,杜鹢,不确定性人工智能,北京:国防工业出版社, 2005:57-185.
[24]张师超,严小卫,王成名,不确定性推理技术,桂林:广西师范大学出版社, 1996:1-49.
[25]李凡,人工智能中的不确定性,北京:气象出版社, 1992: 119-133.
[26]黄昆仑,许文涛,转基因食品安全评价与检测技术,北京:科学出版社, 2009(03).
[27]一民,转基因食品:天使还是魔鬼,中国人民大学出版社, 2010.
[28]魏伟,钱迎倩,转基因生物安全吗,中国林业出版社, 2006.
[29] Lewis Christopher P, Newell James N, Herron Caroline M, et al,Tanzanian farmers’knowledge andattitudes to GM biotechnology and the potential use of GM crops to provide improved levels of foodsecurity. A Qualitative Study,BMC public health, Vol. 10, No.407, Jul. 2010.
[30] Kimenju SC, De Groote H,Karugia J, et al,Consumer awareness and attitudes toward GM foods inKenya,African journal of biotechnology,Vol. 4, NO. 10, pp.1066-1075, Oct. 2005.
[31] Tirado Reyes, Johnston Paul, Food Security: GM Crops Threaten Biodiversity, Science, Vol. 328,NO. 5975, pp.170-171, Apr. 2010.
[32]周彦喆,基于HHM的工程项目风险评价与控制,中国地质大学, 2009.
[33] Haimes, Yacov Y., Hierarchical Holographic Modeling, IEEE Trans Syst man cybernet, Vol. 11,NO. 9, pp.606-617, 1981.
[34] Stan Kaplan, Yacov Y. Haimes, B. John Garrick, Fitting Hierarchical Holographic Modeling intothe Theory of Scenario Structuring and a Resulting Refinement to the Quantitative Definition of risk,Risk analysis, Vol. 21, NO. 5, PP.807-815, Oct. 2001.
[35] Tian-Shy Liou, Mao-Jiun J. Wang, Ranking fuzzy numbers with integral value, Vol. 50, NO. 3, pp.247-255, Sep. 1992.
[36] Kaufman A, Gupta M M, Introduction To Fuzzy Arithmetic: Theory and Application, New York:Van Nostrand Reinhold, 1985.
[37] Arthur P. Dempster, Upper and Lower Probabilities Induced by a Multivalued Mapping, Annals ofmathematical statistics, Vol. 38, pp.325-339, 1967.
[38] G.Shafer, A Mathematical Theory of Evidence, Princeton, NJ: Princeton University Press, 1967.
[39]周光中,基于D-S证据理论的科学基因立项评估问题研究,合肥工业大学, 2009.
[40] Thomas L. Saaty, Modeling unstructured decision problems―the theory of analytical hierarchies,Mathematics and Computers in Simulation, Vol. 20, NO. 3, pp.147-158, Sep. 1978.
[41]刘新宪,朱道立,选择与判断-AHP(层次分析法)决策,上海:上海科学普及出版社, 1990.
[42] Ronald R. Yager,On the dempster-shafer framework and new combination rules, Information Sci-ences, Vol. 41, NO. 2, pp.93-137, Mar. 1987.
[43]张军,涂国平,加权平均法解决证据理论中的失效问题,微计算机信息, 2007, 23(3): 202-203,225.
[44]孙全,叶秀清,顾伟康,一种新的基于证据理论的合成公式,电子学报, 2000, 28(8): 117-119.
[45] Smets P, Belief functions: The disjunctive rule of combination and the generalized Bayesian theorem,International Journal of Approximate Reasoning, Vol. 9, NO. 1, pp.1-35, Aug. 1993.
[46] Smets P, Belief functions on real numbers, International Journal of Approximate Reasoning, Vol. 40,NO. 3, PP.181-223, Nov. 2005.
[47] Haenni R, Are alternatives to Dempster’s rule of combination alternatives?, Information Fusion,Vol. 3, pp.237-241, Jun. 2002.
[48] Haenni R, Lehmann N, Probabilistic argumentation systems: a new perspective on the dempster-shafer theory, International Journal of Intelligent System, Vol. 18, NO. 1, PP.93–106, Jan. 2003.
[49] Haenni R, Shedding new light on Zadeh’s criticism of Dempster’s rule of combination, 8th Interna-tional Conference on Information Fusion, pp. 25-28, Jul. 2005.
[50] Lefevrea E, Colota O, Vannoorenbergheb P, Belief function combination and con?ict management,Information Fusion, Vol. 3, NO. 2, pp.149-162, Jun. 2002.
[51] Smets P, The combination of evidence in the transferable belief model, IEEE Transactions on PatternAnalysis and Machine Intelligence, Vol. 12, NO. 5, PP.447 - 458, 1990.
[52] Zhang Shanying, Pan Quan, Zhang Hongcai, A New Kind of Combination Ruleof Evidence Theory,Control and Decision, 2000, 15(5):540-544.
[53] Pan Quan, Zhang Shanying, Cheng Yong Mei, Zhang Hongcai, Some Research on Robustness ofEvidence Theory, Acta Automatica Sinica, 2001, 27(6): 798-805.
[54] Murphy C K, Combining belief functions when evidence con?icts, Decision Support Systems, Vol.29, NO. 1,pp.1-9, Jul. 2000.
[55]邓勇,施文康,朱振富,一种有效处理冲突证据的组合方法,红外与毫米波学报, Vol. 3, NO. 2,pp.27-32, 2004.
[56] Smets P,Constructing the Pignistic Probability Function in a Context of Uncertainty, UAI’89 Pro-ceedings of the Fifth Annual Conference on Uncertainty in Artificial Intelligence, Vol. 5, pp.29-39,1990.
[2] D.L. Hall, J. Llinas, An introduction to Multisensor Data Fusion, Proc of the IEEE, pp.6-23, 1997.
[3] E. L. Waltz, Data Fusion for C3I, PaloAlto,CA:EWCommunications, pp.217-226, 1986.
[4] W.S. Gesing, D.B. Reid, An integrated multisensor aircraft track recovery system for remote sensing,IEEETrans.Automat.Contc, Vol. 28, No. 3, pp.356-363, 1983.
[5] S.A. Shafer, A. Stentz, C.E. Thorpe, Architecture for sensory fusion in amobile robot, Proc IEEEInt. Conf. Robotics and Automat, SanFrancisco, pp.2002-2011, 1986.
[6] D.J. Kriegman, l.E. Triend, T.O. Binfork,Stereo vision and navigation in buildings for mobile robots,IEEETrans.Robot.Automat, Vol. 5, No. 6, pp.792-803, 1989.
[7] K. Demirbas,Integration of tactile sensors and machine vision for control of robotic manipulators,Robots 9 Conf. Proc, pp. 37-71, 1987.
[8] Konig, A ,et al,Assessment of the safety of foods derived from genetically modified (GM) crops, Foodand chemical toxicology, Vol. 42, NO. 7, pp.1047-1088, 2004.
[9] Engel KH, Frenzel T, Miller A,Current and future benefits from the use of GM technology in foodproduction,toxicology letters,Vol. 127, NO. 1-3, pp.329-336, Feb. 2002.
[10] Zdunczyk Z,In vivo experiments on the safety evaluation of GM components of feeds andfoods,Journai of animal and feed sciencies,Vol. 10, NO. Supplement 1, pp.195-210, 2001.
[11] Beachy Roger N,Current status and future prospects of virus-resistant GM plants in strengtheningfood security,Indian journal of virology, Vol. 19, NO. 1, pp.51-51, Jan. 2008.
[12] La Mura Maurizio, Allnutt Theodore R., Greenland Andrew, et al, Application of QUIZ for GMquantification in food, Food Chemistry, Vol. 125, NO. 4, pp.1340-1344, Apr. 2011.
[13] Nielsen Annika Porsborg, Lassen Jesper, Sandoe Peter, Public participation: democratic ideal orpragmatic tool? The cases of GM foods and functional foods, Public understanding of science, Vol.20, NO. 2, pp.163-178, Mar. 2011.
[14] Cook AJ, Kerr GN, Moore K, Attitudes and intentions towards purchasing GM food, Journal ofeconomic psychology, Vol. 23, No. 5, pp.557-572, Oct. 2002.
[15] Gaskell G, Allum N, Wagner W; et al., GM foods and the misperception of risk perception, Riskanalysis, Vol. 24, No. 1, pp.185-194, Feb. 2004.
[16]金敬强,武富春,信息融合技术的发展现状与展望,电脑与应用, 2006: 50-52.
[17] Varshney PK, Multisensor data fusion, Electronic and communication engineering journal, Vol. 9,NO. 6, pp.245-253, Dec. 1997.
[18] Benaskeur Abder Rezak, Rheaume Francois, Adaptive data fusion and sensor management for mil-itary applications, Aerospace science and technology, Vol. 11, NO. 4, pp.327-338, May. 2007.
[19] Putvinskaya N, Bulasheva N, Cole G, et al., Information technology systems for fusion industry anditer project, Fosion science and technology, Vol. 47, NO. 3, pp.806-813, Apr. 2005.
[20] Lou KN, Lin CJ, An intelligent sensor fusion system for tool monitoring on a machining centre,International journal of advanced manufacturing technology, Vol. 13, NO. 8, pp.556-565, 1997.
[21] Fang YJ, Masaki I, Horn B, Depth-based target segmentation for intelligent vehicles: Fusion ofradar and binocular stereo, IEEE Transactions on intelligent transportation systems, Vol. 3, NO. 3,pp.196-202, Sep. 2002.
[22] Waxman AM, Fay DA, Rhodes BJ, et al., Information fusion for image analysis: Geospatial founda-tions for higher-level fusion , Proceeding of the fifth international conference on information fusion,Vol. 1, pp.562-569, 2002.
[23]李德毅,杜鹢,不确定性人工智能,北京:国防工业出版社, 2005:57-185.
[24]张师超,严小卫,王成名,不确定性推理技术,桂林:广西师范大学出版社, 1996:1-49.
[25]李凡,人工智能中的不确定性,北京:气象出版社, 1992: 119-133.
[26]黄昆仑,许文涛,转基因食品安全评价与检测技术,北京:科学出版社, 2009(03).
[27]一民,转基因食品:天使还是魔鬼,中国人民大学出版社, 2010.
[28]魏伟,钱迎倩,转基因生物安全吗,中国林业出版社, 2006.
[29] Lewis Christopher P, Newell James N, Herron Caroline M, et al,Tanzanian farmers’knowledge andattitudes to GM biotechnology and the potential use of GM crops to provide improved levels of foodsecurity. A Qualitative Study,BMC public health, Vol. 10, No.407, Jul. 2010.
[30] Kimenju SC, De Groote H,Karugia J, et al,Consumer awareness and attitudes toward GM foods inKenya,African journal of biotechnology,Vol. 4, NO. 10, pp.1066-1075, Oct. 2005.
[31] Tirado Reyes, Johnston Paul, Food Security: GM Crops Threaten Biodiversity, Science, Vol. 328,NO. 5975, pp.170-171, Apr. 2010.
[32]周彦喆,基于HHM的工程项目风险评价与控制,中国地质大学, 2009.
[33] Haimes, Yacov Y., Hierarchical Holographic Modeling, IEEE Trans Syst man cybernet, Vol. 11,NO. 9, pp.606-617, 1981.
[34] Stan Kaplan, Yacov Y. Haimes, B. John Garrick, Fitting Hierarchical Holographic Modeling intothe Theory of Scenario Structuring and a Resulting Refinement to the Quantitative Definition of risk,Risk analysis, Vol. 21, NO. 5, PP.807-815, Oct. 2001.
[35] Tian-Shy Liou, Mao-Jiun J. Wang, Ranking fuzzy numbers with integral value, Vol. 50, NO. 3, pp.247-255, Sep. 1992.
[36] Kaufman A, Gupta M M, Introduction To Fuzzy Arithmetic: Theory and Application, New York:Van Nostrand Reinhold, 1985.
[37] Arthur P. Dempster, Upper and Lower Probabilities Induced by a Multivalued Mapping, Annals ofmathematical statistics, Vol. 38, pp.325-339, 1967.
[38] G.Shafer, A Mathematical Theory of Evidence, Princeton, NJ: Princeton University Press, 1967.
[39]周光中,基于D-S证据理论的科学基因立项评估问题研究,合肥工业大学, 2009.
[40] Thomas L. Saaty, Modeling unstructured decision problems―the theory of analytical hierarchies,Mathematics and Computers in Simulation, Vol. 20, NO. 3, pp.147-158, Sep. 1978.
[41]刘新宪,朱道立,选择与判断-AHP(层次分析法)决策,上海:上海科学普及出版社, 1990.
[42] Ronald R. Yager,On the dempster-shafer framework and new combination rules, Information Sci-ences, Vol. 41, NO. 2, pp.93-137, Mar. 1987.
[43]张军,涂国平,加权平均法解决证据理论中的失效问题,微计算机信息, 2007, 23(3): 202-203,225.
[44]孙全,叶秀清,顾伟康,一种新的基于证据理论的合成公式,电子学报, 2000, 28(8): 117-119.
[45] Smets P, Belief functions: The disjunctive rule of combination and the generalized Bayesian theorem,International Journal of Approximate Reasoning, Vol. 9, NO. 1, pp.1-35, Aug. 1993.
[46] Smets P, Belief functions on real numbers, International Journal of Approximate Reasoning, Vol. 40,NO. 3, PP.181-223, Nov. 2005.
[47] Haenni R, Are alternatives to Dempster’s rule of combination alternatives?, Information Fusion,Vol. 3, pp.237-241, Jun. 2002.
[48] Haenni R, Lehmann N, Probabilistic argumentation systems: a new perspective on the dempster-shafer theory, International Journal of Intelligent System, Vol. 18, NO. 1, PP.93–106, Jan. 2003.
[49] Haenni R, Shedding new light on Zadeh’s criticism of Dempster’s rule of combination, 8th Interna-tional Conference on Information Fusion, pp. 25-28, Jul. 2005.
[50] Lefevrea E, Colota O, Vannoorenbergheb P, Belief function combination and con?ict management,Information Fusion, Vol. 3, NO. 2, pp.149-162, Jun. 2002.
[51] Smets P, The combination of evidence in the transferable belief model, IEEE Transactions on PatternAnalysis and Machine Intelligence, Vol. 12, NO. 5, PP.447 - 458, 1990.
[52] Zhang Shanying, Pan Quan, Zhang Hongcai, A New Kind of Combination Ruleof Evidence Theory,Control and Decision, 2000, 15(5):540-544.
[53] Pan Quan, Zhang Shanying, Cheng Yong Mei, Zhang Hongcai, Some Research on Robustness ofEvidence Theory, Acta Automatica Sinica, 2001, 27(6): 798-805.
[54] Murphy C K, Combining belief functions when evidence con?icts, Decision Support Systems, Vol.29, NO. 1,pp.1-9, Jul. 2000.
[55]邓勇,施文康,朱振富,一种有效处理冲突证据的组合方法,红外与毫米波学报, Vol. 3, NO. 2,pp.27-32, 2004.
[56] Smets P,Constructing the Pignistic Probability Function in a Context of Uncertainty, UAI’89 Pro-ceedings of the Fifth Annual Conference on Uncertainty in Artificial Intelligence, Vol. 5, pp.29-39,1990.