CHEMICAL ENGINEERINGTRANSACTIONS VOL. 51, 2016 A publication of The Italian Association of Chemical Engineering Online at www.aidic.it/cet Guest Editors:Tichun Wang, Hongyang Zhang, Lei Tian Copyright © 2016, AIDIC Servizi S.r.l., ISBN978-88-95608-43-3; ISSN 2283-9216 Fault Diagnosis Research of Computer Software Based on the Fuzzy Theory Zhixue Li Weifang University of Science and Technology, Shouguang,Shandong, CHINA lzxkong@163.com With the popularity of computers, all kinds of computer software become indispensable to people's life. The main causes that fault occurs in computer software are that software non-compatibility, illegal operation, error operating, irrational software setup, computer prevention. And there are six symptoms. Using the phase- divided dynamic adjustment weight coefficient method combined with the data inspected by computer software, expert experience, and mechanism analysis to determine the fuzzy diagnosis matrix, the symptoms and causes are established fuzzy corresponding relationship, the fuzzy model of computer software fault diagnosis is established, through the fault symptoms of computer software to judge its fault causes so as to handle faults quickly and ensure the normal operation of computer software. 1. Introduction With the development of society and the popularity of computers, there appears more and more computer software, which develops rapidly. Computer software has been integrated into every aspect of people's daily life, such as entertainment, work, life, and etc. There are more and more companies of which the development of computer software is taken as the main business. Computer languages become diverse. Program development is classified as the compulsory course of current college students. Therefore, it is very necessary to inspect computer software, collect data and diagnose its faults in order to guarantee the normal safe operation of computer software, make computers compatible with the software, and improve the operation efficiency of computer software. Computer software refers to the programs and documents set up in computer system. The programs are computing the processing objects of the tasks and dealing with the description of the rules. Documents are interpretation data that is convenient to know about its program. The program couldn’t work until they are set up into the computer while documents are generally to show people, not necessarily setting up into the computer. The fault diagnosis of computer software is to establish an appropriate model (Cui, 2014) using the study of the data relationship between computer software fault and the symptoms, and then according to its model to diagnose the fault. Because the computer system is a very complex system, its running state is inseparable from the computer software. Once there appears faults in computer software, the computer breakdown may occur, which would affect the normal work of the computer. 2. The fault diagnosis theory of computer software 2.1 The fuzziness and complexity of computer software fault The normal operation of the computer is influenced by computer network, computer environment, computer software and some other conditions. The variation of these factors forms interference to the normal operating computers. At the same time, the normal operation of computer software is also influenced by its running time, load variation, medium variation, and temperature changes. It’s running state and fault factors are fuzzy (Bu and Bu, 2012). The possible causes of computer software fault are software non-compatibility, illegal operation, error operating, irrational software setup, computer prevention, which are also uncertain. DOI: 10.3303/CET1651124 Please cite this article as: Li Z.X., 2016, Fault diagnosis research of computer software based on the fuzzy theory, Chemical Engineering Transactions, 51, 739-744 DOI:10.3303/CET1651124 739 When there are faults in computer software, precursors may appear which are fuzzy as well. It is hard to determine the cause exactly. For the whole computer system is very complex, it is interacted, inter-influenced and inter-restricted by its internal parts. 2.2 The fault diagnosis theory of computer software The process of computer software fault diagnosis is similar to the process of traditional Chinese medicine (TCM) checking its patients. Through inspection, auscultation-olfaction, interrogation and palpation method and one's own experience to diagnose patient’s symptoms, according to the fuzzy relationship between the patient's symptoms and the disease, the patient's disease could be determined. The causes of computer software fault could establish fuzzy relationship with its fault symptoms. When there appear faults in computer software, according to its symptoms, the causes of computer software could be diagnosed, so as to clear troubles and make the normal operation of computer software. Through the prophase investigation, using principal component analysis, the five main causes of computer software fault and its six corresponding fault symptoms could be determined. The specific details are shown in table 1. Table 1: the main causes of computer software fault and its corresponding symptoms Computer software fault Corresponding indexes the causes of computer software fault y software non-compatibility y1 illegal operation y2 error operating y3 irrational software setup y4 computer prevention y5 Fault symptoms of computer software x Software can't be installed. x1 Software can be unable to startup normally. x2 Software icons can't be displayed properly. X3 After turning on the software, it would automatically shut down. x4 During the operation of software, it would suddenlyshut down. x5 Software running results cannot be saved. x6 The general process of computer software fault diagnosis is shown in Chart 1. 3. The fuzzy modelof computer software fault diagnosis The method of computer software fault diagnosis is according to the membership function of fuzzy theory and the fuzzy relationship matrix to carry on the fuzzy correspondence between the fault causes of computer software and its fault symptoms so as to realize the computer software fault diagnosis. (1) According to table 1 and Figure 1, the cause set of computer software fault is determined as follows: N={y1, y2, y3, y4, y5}, The fault symptom set induced by the five computer software fault causes is as follows:M={x1, x2, x3, x4, x5, x6}. 740 javascript:showjdsw('showjd_0','j_0') javascript:showjdsw('showjd_4','j_4') (2) Using the method of ranking by dual comparison, the membership function of computer software fault symptoms is determined as μi, according to the membership function to calculate the corresponding membership degree of symptom xi as μi(xi), the fuzzy vector of computer software fault symptoms could be obtained as follows: X={μ1(x1), μ2(x2), μ3(x3), μ4(x4), μ5(x5)，μ6(x6)}; Let fault symptom sample of computer software M bring about by fault cause N and the membership degree be vi(yi), the fuzzy vector of computer software is: Y={v1(y1), v2(y2), v3(y3), v4(y4), v5(y5), v6(y6)}. Figure 1: The flow chart of computer software fault diagnosis (3) Using the phase-divided dynamic adjustment weight coefficient method combined with the data inspected by computer software, expert experience, and mechanism analysis to determine the fuzzy diagnosis matrix (Armando et al., 2012), the detailed process is as follows: First, according to the monitoring data of computer software fault, the membership degree could be determined as aij. According to the expert experience, the membership degree could be determined as bij. According to the mechanism analysis of computer software fault, the membership degree could be determined as cij. (DavoodSamari et al., 2011) Second, according to the running time, working state, fault times of computer software, the weight coefficient of each membership degree is λ1, λ2, λ3, and when λ1, λ2, λ3≥0, λ1+λ2+λ3=1; Finally, the elements of fuzzy diagnosis matrix could be obtained, that is, the final membership degree: rij=λ1aij+λ2bij+λ3cij, The fuzzy diagnosis matrix of computer software fault could be obtained as follows: 11 12 15 21 22 25 6 5 61 62 65 ( ) ij r r r r r r R r r r r               , And rij∈[0,1],i=1,2,…,6;j=1,2,…,5. (4) With the aid of the fuzzy vector of computer software fault symptom, the fuzzy vector of fault cause and the fuzzy diagnosis matrix, the fuzzy relation equation could be established as follows: Y’=XR, After solving the equation, based on the maximum membership principle, the causes of computer software fault could be determined. 4. The simulation application of computer software fault diagnosis 4.1 Using the method of ranking by dual comparison, the membership function of computer software fault cause could be determined The common method of determining the membership function is the method of ranking by dual comparison. When there appear fault symptoms of computer software, it could determine the sequence of fault causes. In this paper, the alternative optima relative method is chosen, that is, when n fault causes of computer software 741 induce the same fault symptom, enough computer professionals carry on the multiple comparisons to these n fault causes at a time, evaluating which one is the most probable cause bringing about the computer software fault symptoms. And according to the results of the multiple records, the sequence of fault causes under this symptom could be determined. Using the alternative optima relative method, 100 professionals carry on the multiple comparisons to five fault causes of certain software. Record the most probable cause that each professional thinks inducing the computer software fault as an experiment. Every one repeats twice, and the experiment is carried on 60 times. At last, sort out the experiment results of these 100 professionals’ 3000-time experiments. The results are shown in table 2. Table 2: the preferred time data results of five computer software fault causes Y y1 y2 y3 y4 y5 Σyi Ranking vi(yi) y1 0 193 111 104 90 498 4 0.599 y2 74 0 81 93 104 352 5 0.424 y3 228 175 0 100 151 654 3 0.787 y4 261 178 136 0 90 665 2 0.800 y5 352 189 115 175 0 831 1 1 From table 2 we could get that through the alternative optima relative method, the fuzzy vector of membership function the fault symptom xj of computer software to the fault cause yi could be determined as follows: Y=(0.599, 0.424, 0.787, 0.800, 1) 4.2 Construct the fuzzy diagnosis matrix of computer software fault According to the monitoring data of computer software fault, the membership degree could be determined as: 6 5 0.4 0.7 0 0.28 0 0.5 0.45 0.35 0 0 0.44 0 0.48 0.45 0.39 ( ) 0 0 0.6 0 0 0 0 0.5 0 0 0.35 0.4 0.5 0.48 0.55 ij A a                       ; According to the expert experience, the membership degree could be determined as 6 5 0.6 0.8 0 0.72 0 0.5 0.55 0.4 0 0 0.56 0 0.52 0.3 0.61 ( ) 0 0 0.4 0 0 0 0 0.5 0 0 0.4 0.6 0.5 0.52 0.45 ij B b                       ; According to the mechanism analysis of computer software fault, the membership degree could be determined as 6 5 1 0 0 1 0 1 1 0 0 0 1 0 1 0 1 ( ) 0 0 1 0 0 0 0 1 0 0 0 1 1 1 1 ij C c                       ; According to the running time, working state, fault times of computer software(Geng and Xu, 2011), the weight coefficient of each membership degree is as follows: λ1=0.4,λ2=0.4, λ3=0.2; 742 javascript:showjdsw('showjd_0','j_0') http://scholar.cnki.net/result.aspx?q=%e4%bd%9c%e8%80%85%3a(Rubo+Geng)&uid=WEEvREcwSlJHSldRa1FhcmlETDFqMlN2VEM5ejJrM0ZHN040VXFIV0ViS3k4WWFzOVJWNGsyaHNyckJPSTVFRU9RPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!!&UID=WEEvREcwSlJHSldRa1FhcmlETDFqMlN2VEM5ejJrM0ZHN040VXFIV0ViS3k4WWFzOVJWNGsyaHNyckJPSTVFRU9RPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!! Finally, according to the formula rij=λ1aij+λ2bij+λ3cij, the fuzzy diagnosis matrix of computer software fault could be obtained as follows: 6 5 0.6 0.4 0 0.98 0 0.8 0.98 0.3 0 0 0.95 0 0.8 0.3 0.98 ( ) 0 0 0.98 0 0 0 0 0.9 0 0 0.3 0.6 0.9 0.98 0.95 ij R r                       . Details are shown in table 3. Table 3: the fuzzy diagnosis matrix of computer software fault fault causes fault symptoms software non- compatibility y1 illegal operation y2 error operating y3 irrational software setup y4 computer preventio n y5 Software can't be installed.x1 0.6 0.4 0 0.98 0 Software can be unable to startup normally. X2 0.8 0.98 0.3 0 0 Software icons can't be displayed properly. X3 0.95 0 0.8 0.3 0.98 After turning on the software, it would automatically shut down. X4 0 0 0.98 0 0 During the operation of software, it would suddenlyshut down. X5 0 0 0.9 0 0 Software running results cannot be saved. X6 0.3 0.6 0.9 0.98 0.95 4.3 The maximum membership degree principle is used to determine the fault causes of computer software When there are 4 symptoms of computer software fault, that is, software can't be installed x1, software icons can't be displayed properly x3, after turning on the software, it would automatically shut down x4, during the operation of software, it would suddenly shut down x5, use number 1 to express that there appear computer software fault symptoms, use number 0 to express that computer software fault symptoms don’t appear. The fault symptom vector could be obtained as follows: X= (1, 0, 1, 1, 1,0). Fuzzy relation equation is established as follows: 0.6 0.4 0 0.98 0 0.8 0.98 0.3 0 0 0.95 0 0.8 0.3 0.98 (1, 0,1,1,1, 0) (0.95, 0.4, 0.98, 0.98, 0.98) 0 0 0.98 0 0 0 0 0.9 0 0 0.3 0.6 0.9 0.98 0.95 Y X R                       . According to the known fuzzy vector of fault cause Y and the solved Y’, the possibility vector of corresponding fault causes could be calculated as follows: (0.599×0.95, 0.424×0.4, 0.787×0.98, 0.800×0.98, 1×0.98) = (0.569,0.169,0.771,0.784,0.98) Based on themaximum membership principle, the fault cause of computer software could be determined as computer prevention y5. 743 javascript:showjdsw('showjd_4','j_4') javascript:showjdsw('showjd_4','j_4') 5. Conclusion According to the fuzzy theory to establish the fault diagnosis model of computer software(Robin, 2004), it is crucial to query the fault causes of computer software quickly, and ensure the normal operation of computer software and computer itself. The model is fast, accurate and timesaving while the deficiency of this model is that the determination of fuzzy diagnosis matrix needs the real-time data and the expert experience so as to ensure more accurate determination of fault causes. Reference Armando C., Roberta C., Tamara M, 2012. Using Fuzzy AHP to manage Intellectual Capital assets: An application to the ICT service industry [J]. Expert Systems With Applications. DOI: 10.1016/j.eswa.2012.12.081 Bu H.B., Bu S.Z, 2012, Two-Layer Fuzzy Comprehensive RSA-ANP-DSS Evaluation Model of Emergency Management Capacity about Enterprise Value Network [J]. Systems Engineering Procedia, 5, pp.93-98 Cui X.H., 2014, king and analytic hierarchy process in the application of the shortage of water resources evaluation [J]. Journal of mathematics practice and understanding, 6, 270-273. Davood S., Hossein A., Kiumars Z., Gholamhossein H., Frank W, 2012, Determining appropriate forestry extension model: Application of AHP in the Zagros area, Iran [J]. Forest Policy and Economics, 15, pp.91- 97. DOI: 10.1016/j.forpol.2011.10.006 Ding K., Song L.Z, 2015, On the Propaganda Network Construction in Colleges and Universities from the Perspective of We-media [J]. Educational Research, 2: 49-56. Donald J.B, 1998, A model for the software engineering component of a computer science curriculum [J].Information and Software Technology, 40(4), pp.195-201, DOI:10.1016/S0950-5849(98)00039-1. Geng R.B., Gang X, 2011, Application of AHP FSE Method in the Network Course Quality Evaluation[J]. Procedia Engineering, 15, pp.4136-4141. DOI: 10.1016/j.proeng.2011.08.776 Gillmor D, 2004, We the Media: Grassroots Journalism by the People, for the People M. New York NY: O’REILLY& ASSOC INC, 3. Han Z.G, 2005, Mathematical Modeling and Its Application [M]. Beijing: Higher Education Press. Li J.T, 2011, The law protects for computer software [J]. Energy Procedia, 11, pp.1245-1249, DOI: 10.1016/j.egypro.2011.10.385. Li T, 2015, The Application of Fuzzy Mathematics in Satisfaction Degree of Customers in Supermarket [J]. Journal of Capital Normal University, 3: 15-18. Piotr J., Slawomir B., Robert B, 2009, Assessing contractor selection criteria weights with fuzzy AHP method application in group decision environment [J]. Automation in Construction, 19(2), pp.120-126. DOI: 10.1016/j.autcon.2009.12.014 Robin H.K, 2004, Learning performance and computer software: an exploration of knowledge transfer[J]. Computers in Human Behavior, 23(1), pp.333-352, DOI: 10.1016/j.chb.2004.10.029. Zhu D.Q., Gu W, 2008, Sensor Fusion in Integrated Circuit Fault Diagnosis Using a Belief Function Model [J].PInternational Journal of Distributed Sensor Networks, 4(3), 247-261. 744 http://scholar.cnki.net/result.aspx?q=%e4%bd%9c%e8%80%85%3a(Robin+H.+Kay)&uid=WEEvREcwSlJHSldRa1Fhalp4K0NjTCtMQkNLVjljSDloT1ZpcE1wMGpodUExRHhqeWVQUHd6UnBCYkdLUjRBVzFRPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!! http://scholar.cnki.net/result.aspx?q=%e4%bd%9c%e8%80%85%3a(Roberta+Costa)&uid=WEEvREcwSlJHSldRa1FhcmlETDFqMlN2VEM5ejJrM0ZHN040VXFIV0ViS3k4WWFzOVJWNGsyaHNyckJPSTVFRU9RPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!!&UID=WEEvREcwSlJHSldRa1FhcmlETDFqMlN2VEM5ejJrM0ZHN040VXFIV0ViS3k4WWFzOVJWNGsyaHNyckJPSTVFRU9RPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!! http://scholar.cnki.net/result.aspx?q=%e4%bd%9c%e8%80%85%3a(Tamara+Menichini)&uid=WEEvREcwSlJHSldRa1FhcmlETDFqMlN2VEM5ejJrM0ZHN040VXFIV0ViS3k4WWFzOVJWNGsyaHNyckJPSTVFRU9RPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!!&UID=WEEvREcwSlJHSldRa1FhcmlETDFqMlN2VEM5ejJrM0ZHN040VXFIV0ViS3k4WWFzOVJWNGsyaHNyckJPSTVFRU9RPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!! http://scholar.cnki.net/result.aspx?q=%e4%bd%9c%e8%80%85%3a(Donald+J+Bagert)&uid=WEEvREcwSlJHSldRa1Fhalp4K0NjTCtMQkNLVjljSDloT1ZpcE1wMGpodUExRHhqeWVQUHd6UnBCYkdLUjRBVzFRPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!!&UID=WEEvREcwSlJHSldRa1Fhalp4K0NjTCtMQkNLVjljSDloT1ZpcE1wMGpodUExRHhqeWVQUHd6UnBCYkdLUjRBVzFRPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!! http://d.scholar.cnki.net/link/doi/SJES_U/SJES13011501721954?UID=WEEvREcwSlJHSldRa1Fhalp4K0NjTCtMQkNLVjljSDloT1ZpcE1wMGpodUExRHhqeWVQUHd6UnBCYkdLUjRBVzFRPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!! http://scholar.cnki.net/result.aspx?q=%e4%bd%9c%e8%80%85%3a(Rubo+Geng)&uid=WEEvREcwSlJHSldRa1FhcmlETDFqMlN2VEM5ejJrM0ZHN040VXFIV0ViS3k4WWFzOVJWNGsyaHNyckJPSTVFRU9RPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!!&UID=WEEvREcwSlJHSldRa1FhcmlETDFqMlN2VEM5ejJrM0ZHN040VXFIV0ViS3k4WWFzOVJWNGsyaHNyckJPSTVFRU9RPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!! http://scholar.cnki.net/result.aspx?q=%e4%bd%9c%e8%80%85%3a(Gang+Xu)&uid=WEEvREcwSlJHSldRa1FhcmlETDFqMlN2VEM5ejJrM0ZHN040VXFIV0ViS3k4WWFzOVJWNGsyaHNyckJPSTVFRU9RPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!!&UID=WEEvREcwSlJHSldRa1FhcmlETDFqMlN2VEM5ejJrM0ZHN040VXFIV0ViS3k4WWFzOVJWNGsyaHNyckJPSTVFRU9RPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!! http://d.scholar.cnki.net/link/doi/SJES_U/SJES13011300069572?UID=WEEvREcwSlJHSldRa1FhaXVleG15Q2hKSUxPdDROOGplMVA2L3Y3SU55bVhLTy8wVHhqZGFRTDZQTllMSUYwQitBPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!! http://scholar.cnki.net/result.aspx?q=%e4%bd%9c%e8%80%85%3a(Jingtian+Li)&uid=WEEvREcwSlJHSldRa1Fhalp4K0NjTCtMQkNLVjljSDloT1ZpcE1wMGpodUExRHhqeWVQUHd6UnBCYkdLUjRBVzFRPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!!&UID=WEEvREcwSlJHSldRa1Fhalp4K0NjTCtMQkNLVjljSDloT1ZpcE1wMGpodUExRHhqeWVQUHd6UnBCYkdLUjRBVzFRPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!! http://d.scholar.cnki.net/link/doi/SJES_U/SJES13011300096370?UID=WEEvREcwSlJHSldRa1Fhalp4K0NjTCtMQkNLVjljSDloT1ZpcE1wMGpodUExRHhqeWVQUHd6UnBCYkdLUjRBVzFRPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!! http://scholar.cnki.net/result.aspx?q=%e4%bd%9c%e8%80%85%3a(Piotr+Jaskowski)&uid=WEEvREcwSlJHSldRa1FhaXVleG15Q2hKSUxPdDROOGplMVA2L3Y3SU55bVhLTy8wVHhqZGFRTDZQTllMSUYwQitBPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!!&UID=WEEvREcwSlJHSldRa1FhaXVleG15Q2hKSUxPdDROOGplMVA2L3Y3SU55bVhLTy8wVHhqZGFRTDZQTllMSUYwQitBPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!! http://scholar.cnki.net/result.aspx?q=%e4%bd%9c%e8%80%85%3a(Slawomir+Biruk)&uid=WEEvREcwSlJHSldRa1FhaXVleG15Q2hKSUxPdDROOGplMVA2L3Y3SU55bVhLTy8wVHhqZGFRTDZQTllMSUYwQitBPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!!&UID=WEEvREcwSlJHSldRa1FhaXVleG15Q2hKSUxPdDROOGplMVA2L3Y3SU55bVhLTy8wVHhqZGFRTDZQTllMSUYwQitBPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!! http://scholar.cnki.net/result.aspx?q=%e4%bd%9c%e8%80%85%3a(Robert+Bucon)&uid=WEEvREcwSlJHSldRa1FhaXVleG15Q2hKSUxPdDROOGplMVA2L3Y3SU55bVhLTy8wVHhqZGFRTDZQTllMSUYwQitBPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!!&UID=WEEvREcwSlJHSldRa1FhaXVleG15Q2hKSUxPdDROOGplMVA2L3Y3SU55bVhLTy8wVHhqZGFRTDZQTllMSUYwQitBPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!! http://d.scholar.cnki.net/link/doi/SJES_U/SJES13011501834173?UID=WEEvREcwSlJHSldRa1FhaXVleG15Q2hKSUxPdDROOGplMVA2L3Y3SU55bVhLTy8wVHhqZGFRTDZQTllMSUYwQitBPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!! http://scholar.cnki.net/result.aspx?q=%e4%bd%9c%e8%80%85%3a(Robin+H.+Kay)&uid=WEEvREcwSlJHSldRa1Fhalp4K0NjTCtMQkNLVjljSDloT1ZpcE1wMGpodUExRHhqeWVQUHd6UnBCYkdLUjRBVzFRPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!! http://d.scholar.cnki.net/link/doi/SJES_U/SJES13012100262343?UID=WEEvREcwSlJHSldRa1Fhalp4K0NjTCtMQkNLVjljSDloT1ZpcE1wMGpodUExRHhqeWVQUHd6UnBCYkdLUjRBVzFRPT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!! http://scholar.cnki.net/result.aspx?q=%e4%bd%9c%e8%80%85%3a(Daqi++Zhu)&uid=WEEvREcwSlJHSldRa1FhcmovUHRxOTNwOFVQdUdaVHlzd3hvbG5YVS9ZbVloaWJrNFJWVjFuVEZRNTByZnl3SS93PT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!!&UID=WEEvREcwSlJHSldRa1FhcmovUHRxOTNwOFVQdUdaVHlzd3hvbG5YVS9ZbVloaWJrNFJWVjFuVEZRNTByZnl3SS93PT0=$9A4hF_YAuvQ5obgVAqNKPCYcEjKensW4IQMovwHtwkF4VYPoHbKxJw!!