A novel approach to intelligent scheduling based on fuzzy feature selection and fuzzy classifier

Intelligent scheduling policy is needed for dynamic manufacturing systems, which tailors the dispatching rule adaptively depending on the prevailing state of the system. In this kind of system-features-oriented intelligent scheduling method, appropriately choosing the group of system features is very critical to scheduling performance. However, research efforts on this topic are limited. In this article, based on general fuzzy model and heuristic searching tree, a new quantitative method is proposed to evaluate and reduce the system features. Then, an intelligent scheduler using fuzzy classifier is created, which can present fuzzy or uncertainty cases. The fuzzy classifier is constructed by fuzzy neural network. Tests of simulation and contrast show that, through applying fuzzy logic in intelligent scheduling framework, the learning performance of classifier and scheduling performance of scheduler are both improved. In conclusion, the fuzzy feature selection and fuzzy pattern classifier methods have great potential in improving the performance of intelligent scheduler.