The Research On Search Reasoning And Learning Methods For Automotive Stamping Halting Problems

Automobile industry as the core pillar industry of the national economy, gathered a large number of new and high technology and top talents. In the face of increasingly fierce Global integration market competition environment, improve the core technology and research ability is the key of the automobile enterprise strategy. However, the core technology and research ability rely heavily on the accumulation and effective use of experience knowledge.Stamping process as one of the most important process of vehicle manufacturing technology, its level can have influence directly on the enterprise production efficiency and quality. The state of production line stop during pressing called stamping halting problem in this article. If stamping halting problem can be handled efficiently, it will become an effective mean to improve production efficiency and quality of the whole car. Meantime, workers will accumulate enough knowledge of experience when deal with stamping halting problem, all workers experience knowledge can be formed organization experience knowledge, how to use organization experience knowledge efficiently providing decision support for employees, and learn the knowledge of processing in time are the key points of this study.According to knowledge structure of stamping halting problem and quantity characteristics, this article puts forward the search reasoning and learning methods of stamping halting problem based on processing knowledge system and presentation model of current research results. Through collecting daily workers experience knowledge can provide decision support for employees to handle new stamping halting problem by using search reasoning. It can reuse stamping halting problems knowledge efficiently and improve organization ability and efficiency of solving stamping halting problem. The new knowledge produced during the reasoning process will be added in stamping halting problems knowledge in time, so that ensure real-time, dynamic and correctness of stamping halting problem processing knowledge.For the research of search reasoning methods of stamping halting problem processing knowledge, In analogy to human solving the problem this article provide decision support for employees based on the stamping halting problem processing knowledge network model, combining with heuristic search strategy. In order to meet the demand of employees of different decision, this article design two kinds of methods, online shallow heuristic search and offline deep heuristic search, based on the structural features of stamping halting problem processing knowledge network.For study methods research of stamping halting problem processing knowledge, this article puts forward a learning strategy based on reinforcement learning, and design corresponding learning method for two search reasoning methods, so that can study the knowledge at the process of search reasoning.In the end, this article will construct a stamping halting processing prototype system, regarding the search reasoning and learning method as theory foundation, object-oriented analysis and design as the guiding ideology, JavaScript as the development of language. And choose an enterprise part of actual data in the process of actual production as an example of actual data analysis and system display to verify feasibility of search reasoning and learning method.