Integrating Knowledge Acquisition and Deduction in Product Configuration by Neural Networks

In the development of product configurator, knowledge acquisition and deduction are the primary tasks which extract and represent configuration knowledge and then embody it into configurator to facilitate configuration process. Product configuration is implemented in dynamic environment where customers´ preference and technical solutions always keep changing. Therefore, configuration knowledge should be timely modified to maintain the effectiveness of configuration activities. In traditional manner, knowledge acquisition and deduction are mostly addressed isolatedly. The separated knowledge acquisition and deduction probably result in long response time and involve a great many human and organizational resources to keep configurator competent. To address this problem, this research adopts neural network technique and specific rule extraction mechanism to develop product configurator. The proposed method is able to integrate knowledge acquisition and deduction into a connectionist methodology and meanwhile provide comprehensible knowledge representation for product configuration activities.