Research On Digital Twin Based Edge-cloud Collaborative Production Scheduling

With the wide application of the new generation information technologies in intelligent production,new smart manufacturing mode puts forward higher requirements for real-time and accuracy of production scheduling.The application of digital twin in production scheduling can achieve interaction and mapping between information space and physical space,and improve the flexibility of scheduling strategies,so that the uncertain events in the production process can be better dealt with.For the production scheduling problem in large-scale smart manufacturing environment,it is difficult for the cloud to fully meet the demand of digital twin application for data processing and real-time response capability of the terminals.Therefore,the digital twin based edge-cloud collaborative production scheduling in manufacturing environment with multiple workshops is studied in this thesis.Firstly,the digital twin based edge-cloud collaborative production scheduling scheme is studied.Based on the edge-cloud collaborative production scheduling system,the dynamic production scheduling strategies and optimization methods are designed.This scheme can train neural networks to simulate and predict the production capacity of each workshop.It uses genetic algorithm to normalize the packaging of the tasks from different customer orders and give the allocate the task-packages to each workshop in the cloud,optimizing the allocation of global manufacturing resources.The workshops carry out distributed scheduling on the edge-side to improve tasks processing efficiency.In the production process,the cloud obtains the parameters of the workshops periodically,updates the task-package completion time prediction models of them,and adjusts scheduling plans in real time,responding to the fluctuation of production condition effectively,and ensuring the production efficiency.Simulation results verify the effectiveness of the digital twin based edge-cloud collaborative production scheduling scheme.Aiming at the allocation problem of algorithms tasks in different stages of the scheduling process,this thesis designs the corresponding adaptive algorithms tasks allocation strategy models and allocation optimization methods,adaptively allocating the algorithms tasks in the cloud or the edge-side according to the computing ability.In this thesis,improved genetic algorithm is used to allocate the algorithms tasks of workshop production capacity prediction.Greedy algorithm is combined to design the algorithms tasks allocation method for the flexible job shop scheduling of the task-packages.Compared with the computing locations fixed algorithms tasks allocation method,the adaptive algorithms tasks allocation method can make full use of the computing resources of the edge-cloud collaborative system,reduce the scheduling time and improve the scheduling efficiency.The advantages of the adaptive algorithms tasks allocation strategy are more significant in large-scale production environment or mass order production.