Performance Evaluation Of Discrete Manufacturing Workshops Based On Key Performance Indicators And Neural Network

The workshops are not only the production center of manufacturing enterprises, but also the source of their profits. Workshops are more and more important for the development of manufacturing enterprises. The manufacturing level of workshops directly affects competitive advantages of manufacturing enterprises. Because of the diverse marketing demands, complex manufacturing process, diverse materials and relatively complex management process of discrete manufacturing workshops, it is more and more difficult to improve manufacturing level of workshops. A large number of scholars spoke highly of the positive role of the performance evaluation in process reengineering and continuous improvement. They also consider that manufacturing performance evaluation is helpful for decision-making of corporates and improvement of workshops’management. Performance evaluation helps the corporates find the weaknesses and potential problems in order to realize lean management of workshops. Therefore, it becomes a deeply meaningful research topic to objectively establish discrete manufacturing workshop performance indicator system and to select effective performance evaluation methods.To achieve the effective and scientific evaluation on discrete manufacturing workshop level performance, a key performance indicator double-dimension model of discrete manufacturing workshops was built. Based on the standard ISO 22400 about Key Performance Indicators for Manufacturing Operation Management, Value Tree model was used to break down the strategy objectives of discrete plants vertically, and workshop manufacturing operation management was analyzed horizontally by Fishbone Diagram model. This model achieves double-dimension performance evaluation, one is the vertical evaluation for executing processes of enterprise level strategy goals, the other is the horizontal evaluation for workshop manufacturing operation management. Then the function model for performance evaluation system was built with the semantic data modeling method IDEFO. This model ensures informational realization on performance evaluation architecture of discrete manufacturing workshop. An application example proved this evaluation architecture is feasible and operable. This architecture can provide an effective method on performance evaluation for discrete manufacturing workshop. After researching bionics theory, performance evaluation method of discrete manufacturing workshop was proposed based on BP neural network algorithm, which used linear approximation method to optimize the performance. An application example proved the suggested evaluation method is feasible and operable, which supplies policy making tools for discrete manufacturing workshop.