基于BP神经网络的不确定性动态Job-shop调度研究

Research on Uncertainty Dynamic Job-shop Scheduling Based on BP Neural Networks

针对作业车间调度的不确定和动态问题,围绕设备故障、订单追加、紧急订单插入3种典型的不确定性情况提出1种基于BP神经网络的重调度方法。当生产过程中发生不确定性事件对原先调度方案产生巨大扰动时,通过已构建且训练好的BP神经网络快速进行响应并生成1个重调度方案,保证整个生产过程高效、有序、稳定地运行。通过仿真实例验证了可行性。

The uncertainty exists in job shop scheduling. It is mainly caused by machine fault, additional order and emergency order insertion. This paper proposes a method of the rescheduling based on BP neural network. When the uncertainty event appears in the production process and it has great influence on the original scheduling, the BP neural network already constructed and trained can be used to quickly generate a re-scheduling scheme which is used to make sure that the effective and smooth operation is guaranteed in the whole production process. The feasibility is verified by the simulation example.