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

As a kind of knowledge intensive industry, automotive industry has accumulated a bunch of advanced technologies and processes, so the modern technologies and knowledge innovations have become competitive advantages in place of the original capital in this industry. That’s why the demands for the knowledge management especially in operation layers are increasingly urgent in manufacturing enterprises nowadays.As the first process in automobile production, the work pieces in stamping process directly affect the cost and quality of the automotive products entirely, so the continuous breakdowns in stamping production lines have become the bottleneck in the whole stamping productivity for automotive enterprises. When the workers in production lines are handling these problems, they have accumulated a great deal of empirical knowledge which develops into important knowledge capital for enterprises. In this circumstance, this paper aims to realize effective utilization of this intellectual capital from the perspective of knowledge management and provide practical methods for production management in automotive enterprises.Based on the knowledge system of halting problem system, this paper analyzes the characteristics of knowledge in halting problem processing and put forward relevant reasoning and learning methods. In the uncertainty reasoning of halting problem knowledge, this paper introduces the Causality Structure Diagram (CSD) as knowledge representation model and establishes a hybrid reasoning model combining model-based reasoning and rule-based reasoning. In accordance with one single question reasoning and multiple questions reasoning conditions, two different reasoning mechanisms consists of relevant reasoning rules are set up in this paper.To realize knowledge learning of problem processing, this paper analyzes all kinds of knowledge learning situation generated in reasoning process and set up the parametric and the structural learning methods on CSD to renew the existing knowledge system.On the basis of uncertainty knowledge reasoning and learning method proposed above, this article introduces decision support system as an implementation tool. According to the basic development ideas of decision support system, this paper designs the overall structure of the prototype system and implements the system. At last an example is analyzed by using real production data from a car manufacturing plant to verify the validity and practicability of these reasoning and learning methods.