Research On Fuzzy VFT Modeling And Its Algorithm

There are always a large number of hierarchies in weapons system-of-systems. And the relations between their factors and components are very complex. Therefore, it is very important and challenging to model and evaluate weapons system-of-systems. However, traditional AFT (Alternative Focused Thinking) methods base the decision process only on the existed alternatives, which limits the decision maker's view. In contrast, newly proposed VFT (Alternative Focused Thinking) method takes value as the starting and central point of decision. And it can help to design novel and more preferable alternatives. However, VFT has not considered about fuzziness contained in decision process, especially introduced by human decision maker or participants. Viewing the problem, this thesis proposes the Fuzzy VFT (FVFT) method, which is a fuzzy extension of VFT.In summary, the main works and contributions of this thesis are:(1) It discusses about the necessity of fuzzy extension to VFT. It presents the basic concepts, framework and steps of FVFT. Moreover, it analyzes and points out that value model weighting, alternatives'value computation and ranking should be the focuses of following research.(2) Viewing that the value model has a strict tree-like structure, the thesis suggests to use AHP (Analytical Hierarchy Process) to implement weight extraction. It analyzes the drawbacks of classical AHP and existed FAHP (fuzzy AHP) methods. Based upon this, the thesis proposes a new FAHP method, i.e. PWRB-FAHP (Pairwise Weight Ratio Based Fuzzy AHP) method to weight the fuzzy value model of FVFT. It describes the weighting processes of PWRB-FAHP when using triangular and trapezoidal fuzzy numbers respectively to depict factors'weights.(3) Since triangular fuzzy numbers can be viewed as particular cases of trapezoidal ones, the thesis employs the latter to describe alternatives'values universally when discussing fuzzy weights computation and ranking. It adopts the max-min based fuzzy number ranking method. Viewing that the method is largely a conceptual one and lacks of operability, the thesis designs computer algorithms for it to realize automatic and fast fuzzy value ranking, which is very essential especially for the existence of a large number of alternatives.(4) Finally, the thesis introduces two decision problems to verify the effectiveness and rationality of proposed FVFT. The first problem is the Mexico City Airport Citing, which is comparatively simple and originally used in some literatures addressing decion problems containing uncertainty. Comparison of the analysis results obtained throught FVFT and those given in other literatures displays the effectiveness of FVFT for handling decision problems with uncertainty. The second problem is the configuration alternatives evaluation of field aerial defence weapon system, which is more practical and complex. The decision results show the effectiveness and rationality of FVFT for large-scale, complex decision problems.