Developing Support System in Fuzzy Control/Decision

• Main Authors: Chy Chyang Chou, 周志強
• Other Authors: Hung Chang Lee
• Format: Others
• Language: zh-TW
• Published: 1994
• Online Access: http://ndltd.ncl.edu.tw/handle/03343204057559277501

碩士 === 淡江大學 === 資訊管理研究所 === 82 === A Fuzzy Control/Decision Support System (FCDS) is presented in this thesis which includes some basic functionality, such as multi-choice fuzzy implication functions, sentence connection logic, fuzzification-defuzzification strategies and build-in membership functions. Additional features about FCDS are meta- level fuzzy inference, soft-cluster rule generator, image support interface. For the meta-level fuzzy inference, it adopted from traditional expert system, and allowed a higher level of reasoning to inference the lower level of the IF-THEN control rules. Likewise, meta-level is the control of information as its traditional meaning. However, we add the fuzzing in the procedure of controlling information and thus the overall reasoning acts more like the meta-level conception in human being. The soft-cluster rule generator generate the rule automatically from the trained data. It will cluster the coming data automatically or interactively to build the degree of the belongings, which is similar to the procedure of building the linguistic variables and the membership function in fuzzy system. The image support interface offers some adjustabilities in the enhancement of the image to recover from ill-posed image, e.g., shielding and shadowing. Finally, we take several kinds of application to demonstrate the capability of FCDS. They include "truck backer-upper control problem", "discuss the logicality between implication function and sentence connection", "image enhancement adjustability", "discern the surface of the road way which is good or bad", and "diagnose the problem of financial liquidity".