Bp Network-based Product Modeling Evaluation System

Product model evaluation is to make final decision solve complex problems that have more solutions. Multi-Target evaluation has already become an important step to improve design quality for product model.The paper summarizes evaluation factors about industrial product model from the points of things and man, and sets up a relatively perfect evaluation target system facing design phases about product model, which becomes a main step for the final decision-making. Selecting appropriate evaluation method is significant to the correctness of the evaluation results. Product model evaluation is characterized by the diversity of evaluation targets, intuitiveness of the judgment, the relativity of evaluation results, and so on. Conventional evaluation methods are more likely to rely on man-made intervention for defining each evaluation target weight. And ANN has self-culture ability. By learn samples, it can obtain a function of weight harmony to each evaluation target weight, which can get rid of as much as possible the man-made intervention in weight defining, insure the validity of the weights defined. This method is useful to weaken man-made intervention in confirming each evaluation target weight. However, this evaluation method based on ANN needs the results from conventional evaluation to train network sample patterns, so this evaluation system adopts a combined evaluation method of grey theory and BP network as product model evaluation. The evaluation system applies Access Database for data management.The system uses Visual C++6.0 for programming and can obtain better design objects. In the paper, the author introduces in detail the method to use the system and verifies the effectiveness of the system by practical examples.