The Research Of Knowledge-Based Real Time Scheduling Support Platform

With the increasing competitiveness of the manufacturing industry, how to obtain a satisfied scheduling sequence in reasonable time becomes a key factor in production efficiency. On the other hand, during the actual production process, the production state is so complex and unpredictable that the advanced scheduling sequence may lose its function after some unknown incidents. How to respond quickly to those changes and re-schedule has become the focus of the study of production scheduling.In this thesis a scheduling support system based on LPKI model is presented including CBS, RBS, MBS and other modules. Artificial neural network is used for an inference engine based on the function of learning and memory of it and an adaptive rule-based scheduling model (RBS) is established. What's more, the representation and reasoning of determinist scheduling problems are studied and a new kind of case-based scheduling (CBS) is presented. Finally, a real time scheduling platform is established based on RBS and CBS in the environment of VC++6.0. The major contributions of this thesis can be summarized as follows.1. An adaptive rule-based scheduling model (RBS) is set with the artificial neural network as the inference engine. This model can respond quickly to the changes in production process and compute instantly to adapt to the real-time requirements of dynamic scheduling. The inputs and outputs of artificial neural network are the state parameters of the manufacturing workshop and the current scheduling rules to be used.2. A case-based scheduling (CBS) method is represented. The scheduling problems are represented by a model in deterministic scheduling theory which consists of 3 objects and its reasoning mechanism is based on the computational complexity in scheduling. The knowledge representation and reasoning are closely integrated with scheduling theory, the way of which makes modeling simple and convenient.3. The framework of knowledge-based production scheduling platform is established, which consists of the mentioned modules RBS and CBS, then the workflow and database are designed, finally computing algorithm for searching cases in CBS module is given.