Microsoft Word - 6 Bing Zhang, Ying Wang, Dejun Chen--A Research on Technology Project Credit Evaluation Model Based on AHP and Advances in Systems Science and Applications (2011), Vol.11, No.3-4 249-256 ISSN 1078-6236 International Institute for General Systems Studies, Inc A Research on Technology Project Credit Evaluation Model Based on AHP and FCEM Bing Zhang, Ying Wang and Dejun Chen School of Information Engineering, Wuhan University of Technology, Wuhan 430070, P.R.China Abstract This paper analyzes the basic requirements of the technology project credit evaluation, and presents a technology project credit evaluation model based on Analytic Hierarchy Process (AHP) and Fuzzy Comprehensive Evaluation Method (FCEM). Combining with the actual situation of one scientific research center, a technology project credit evaluation index system is established, and its weight of each evaluation index is determined by AHP, and then an evaluation results is analyzed and evaluated through FCEM, also it’s valuable theoretical foundation for the management of the Technology Project. Keywords AHP; FCEM; Project Credit Evaluation 1. Introduction The technology project credit evaluation is an important part of technology project process management. In the technology project concluding, it is necessary to evaluate the technology project implementation process, which is not only the summary of project implementation process, but also the archive of credit of undertakers, which provide important historical basis for the future project application approval procedures. Currently, the technology project credit evaluation is based on subjective qualitative assessment, there is no reasonable technology project credit evaluation index system, or the evaluation indexes are lack of scientific weight distribution. To solve this problem, this paper presents a comprehensive evaluation method based on AHP and FCEM. Combining with the actual situation of one scientific research center, this paper proposes a technology project credit evaluation index system, which evaluate the credit of project stakeholders from the contract compliance, reporting significant matters, and implementation within the stipulated time those three aspects, and determines the weight of each index by AHP and checks the consistency, then taking one credit evaluation results in a project as example, calculates the technology project credit situation through fuzzy analysis and quantitative assessment to validate this model. 2. Technology project credit evaluation index system Technology Project credit evaluation system should be operated from the multi-level, multi-angle, which should be able to fully reflect the technology project's credit rating commitment, combined with an actual situation of R & D center, based on AHP, a three-level technology project credit evaluation index system is proposed , as shown in Figure 1. Figure 1 show that, this system is made up of 3 respects of contract compliance, reporting significant events and implementation within the stipulated time, and has 8 indexes; of course, it can be adjusted according to actual situation. Explain the specific content of each index as follows: (1) Completion condition of assessment indicators: refers to the completion condition of the content stipulated in the contract. (2) Rate of progress is the completion situation of the project progress according to the contract rules. (3) Reporting significant events is to account for the significant issues to the virtual 250 Zhang: A Research on Technology Project Credit Evaluation Model Based on AHP and FCEM coordination center faithfully and timely. U 1U 2U 3U 11U 12U 31U 32U 32U 34U 35U Figure 1 Technology project credit evaluation index system (4) Submit research plan: refers to submit their work outline within the specified time, such as it is finished after being noticed in two months. (5) Submit contract: refers to hand over contract within the specified time, such as it is finished after being noticed in three months. (6) Submit the sheet about the execution situation: refers to submit it in scheduled time, such as finishing it on 15 of the first month of each quarter. (7) Submit the acceptance of applications: refers to submit it before the deadline specified in the contract. (8) Submit archive data: refers to submit it within the specified time after project acceptance, such as 1 month after the inspection (30 days) for submission. 3 The Establishment of Technology Project Credit Evaluation Model Based on the AHP and FCEM AHP is used first to define the weight of each level index, and then use FCEM for project credit evaluation. The model is established in following steps: 1.tablishment of factors set From the project credit evaluation index system in Figure 1 can be seen, there are two levels of evaluation indices, and now the first level is defined },,{ 321 UUUU = ; the second level is },{ 12111 UUU = ， },,,,{ 35343332313 UUUUUU = 。 2. Establishment of reviews set According to the actual situation of the R&D center mentioned before, reviews are set to 4 levels, namely, n=4, },,,{ 4321 VVVVV = represent {excellent, good, medium, poor}. Reviews level can be adjusted and determined according to the specific situations. 3. Determine weights set of evaluation indexes by AHP Using the AHP to determine weights set of evaluation indexes can be divided into the following steps: (1) Construct Judgment Matrix Judgment Matrix represents the relative importance between two elements in the same level to some element in the upper level, the evaluation about the importance of indexes at all levels are a subjective process, based on expert evaluation results or the result of the questionnaire. Use Advances in Systems Science and Applications (2011), Vol.11, No.3-4 251 ISSN 1078-6236 International Institute for General Systems Studies, Inc ),...2,1,(, njibb ji = to represent the indexes. ijb expressed the value of the importance that ib relative to jb , construct the judgment matrix P through 1-9 ratio scaling, and the matrix have reciprocity and basic consistency, that is, 0>ijb ， 1=iib ， 1=∗ jiij bb .   ⎥ ⎥ ⎥ ⎥ ⎦ ⎤ ⎢ ⎢ ⎢ ⎢ ⎣ ⎡ = nnnn n n bbb bbb bbb p 21 22221 11211                                       (1)                      Table 1 1-9 Description of Proportion Quotients ijb Meaning Explanation 1 equal importance Both have the same importance 3 Weak Importance ib important than jb slightly 5 Strong Importance ib important than jb obvious 7 Very Strong Importance ib more important than jb 9 Absolute Importance ib absolute important than jb obvious 2、4、6、8 Between the various levels above The importance is between the adjacent levels 1、1/2、… 、1/9 Reverse comparison The importance of jb relative to ib With experience and knowledge related to credit evaluation and project management, construct judgment matrix of each level as 1P 2P 3P . In order to facilitate analysis and more intuitive, we graphically shows the matrix, as shown in Table 2, 3, 4. Table 2 Judgment Matrix 1P Table 3 Judgment Matrix 2P Table 4 Judgment Matrix 3P 3P 31U 32U 33U 34U 35U 31U 1 1 1 1/2 2 2P 11U 12U 11U 1 2 12U 1/2 1 1P 1U 2U 3U 1U 1 2 1/2 2U 1/2 1 1/3 3U 2 3 1 252 Zhang: A Research on Technology Project Credit Evaluation Model Based on AHP and FCEM 32U 1 1 1 1/2 2 33U 1 1 1 1/2 2 34U 2 2 2 1 3 35U 1/2 1/2 1/2 1/3 1 (2) Calculate the weight vector of each level and make consistency check First, calculate the maximized eigenvalue and eigenvector of judgment matrix. Generally, it is to use geometric averaging (root method) or normative column average (sum method) to calculate the approximate eigenvectors [2], and then calculate the Maximized Eigenvalue. Geometric average method: Calculate the product of each element of each row, then calculate the nth root of each product; and then normalized the obtained vector. Vector obtained above is the approximate eigenvectors, if the consistency check is passed, the vector is the relative weight vector of each index. Calculating the maximized eigenvalue and eigenvector using geometric averaging mean is as follows: ① Calculate the geometric average of all elements of each row of the judgment matrix. Based on n n j iji bw ∏ = = 1 ， ( )ni ,,2,1= ，so ( )Tnwwww ,,, 21= 。 ② Normalize w , ∑ = = n i iii www 1 ， ni ,,2,1= , then T ni wwwW ),,,( 2= which is approximate eigenvectors, and its value of each element is the weight value of each index. ③ Calculate the largest eigenvalue maxλ ( )∑ = = n i i i nw PW 1 maxλ (2) In the formula 2, vector ( )iPW is the first i component of PW . Then check on the consistency of judging matrix. Matrix consistency test as follows: ① Calculate the inconsistent level (CI) of Judgment Matrix. 1 max − − = n n CI λ (3) In the formula 3, maxλ is the Maximized eigenvalue of )1( >nn order matrix. ② Calculate the Random Consistency level (RI) of Judgment Matrix, which only determined by the order of the judgment matrix. Note that, when 20 ≤< n , there is no inconsistency issue, matrix does not need be tested. Standards of Random consistency level shown in table 5: ③ Calculate the consistency ratio of judgment matrix (CR). RI CICR = (4) Advances in Systems Science and Applications (2011), Vol.11, No.3-4 253 ISSN 1078-6236 International Institute for General Systems Studies, Inc Table 5 Random consistency level RI n 1 2 3 4 5 6 7 8 9 10 RI 0 0 0.58 0.9 1.12 1.24 1.32 1.41 1.45 1.49 The method to determine the consistency of judgment matrix is [3]: when .10