Multi-strategies in Case Based Design Knowledge Management on Web

Design knowledge management is very important to improve the design process performance. The case based design (CBD) is a methodology to follow mankind's problem solving strategy, one of its aims is to realize the design knowledge reuse. This method has been used in many fields, but there are still some difficulties that hindered its application. The first is about the difficulty in the formalization of a design case, especially representation of some experiences that was the weakness in former researches. the second is about the bottleneck of knowledgeacquisition that results in lacking enough knowledge in the modification process, which may lead to unreliable results when automatic inference is applied. We proposed that modification should be done not by machines but by designers while the inference results can be used as a suggestion to designers. In this manner we integrated both case-based and rule-based methodologies together. The third difficulty relates to case evolution and customization, former CBD systems suffered from this problem. We proposed an agilecase-template technique that overcomes the problem perfectly. To make users be able to define their own similarity rules we allow users to choose different parameter weights and fuzzy subjection functions of numeric parameters. The approach of this paper has been applied to an enterprise and the system is built on the Web.

Design knowledge management is very important toimprove the design process performance. The casebased design (CBD) is a methodology to followmankind's problem solving strategy, one of its aims isto realize the design knowledge reuse. This methodhas been used in many fields, but there are still somedifficulties that hindered its application. The first isabout the difficulty in the formalization of a designcase, especially representation of some experiencesthat was the weakness in former researches. The second is about the bottleneck of knowledge-acqulsition that results in lacking enough knowledge in the modification process.which may lead to unreliable results when automatic inference is applied.We proposed that modification should be done not by machines but by designers while the inference results can be used as a suggestion to designers.In this manner we integrated both case.based and rule-based methodologies together.The third difficulty relates to case evolution and customization.former CBD systems suffered from this problem.We proposed an agile case-template technique that overcomes the problem perfectly.To make users be able to define their own similarity rules we allow users to choose different parameter weights and fuzzy subjection functions of numeric parameters.The approach of this paper has been applied to an enterprise and the system is built on the Web.