Complex Product Performance Configuration Design On Extension Classification Knowledge Mining

As one of several essential technologies for mass customization, product configuration design can tailor individual products quickly and has been adopted as one important and effective means for contemporary enterprises to participate in fierce market competition. After taking home and aboard existing methods into insight, the paper investigates contradictory problems between performance primitives and existing structure in conjunction with extension logic theory, cased based reasoning (CBR) and extension classification mining et al. It explores one formal and intelligent design method, which can be of great importance to theory research and engineering applications.1. Taking expression of configuration knowledge and depth of hierarchical structure in product configuration models into account, extension theory has basic and compound elements as tools to analyze constraint types of product components, and to establish formal models of product existing cases and that of product configuration reasoning. It can expand its applicable scope appropriately.2. CBR can be utilized to assist product performance configuration design, and fundamental characteristics of interval to interval similarity in CBR analyzed. Interval to interval similarity is established according to improved distance. Numeric to interval and numeric to numeric similarity are corresponding special cases of the formal. Results indicate that the method has higher distinction and can meet fundamental characteristics of similarity perfectly. It also can deal with indeterminate, fuzzy customer performance primitives effectively to apply to a wider range of diverse environments.3. In the first search through product cases, extension classification mining method is applied to extract static design knowledge. In the process of product configuration design, a basic transformative library is established through divergence analysis. Then operators are randomly generated to perform some operations on existing structure, and transformative directions controlled to be in accordance with objectives. Furthermore, transformative magnitude is computed to weigh its ease of operation. Finally, dynamic, qualitative or quantitative transformative design knowledge is extracted.4. Motor and fore handler in the new chain saw configuration are simulated, and it is in compliance with the trend of theoretical results. Finally, the original system is developed to validate feasibility of the method above in the paper and applied to solve contradictory problems intelligently in the product performance configuration design.