Research On Flexible Job Shop Scheduling Based On Digital Twin

As the economic development enters the new normal stage,Chinese manufacturing industry is also actively transforming to a quality and efficient manufacturing mode.At the same time,with the continuous promotion of the digital economy,the production characteristics of the manufacturing industry have also undergone tremendous changes.The manufacturing mode has shifted to a multi-type,small-scale manufacturing mode,and this will make the manufacturing process more complicated and too difficult to be controlled.As an important part of the manufacturing process,intelligent and automated scheduling is of great significance for the enterprise to regain competitiveness.The research goal of intelligent scheduling is to achieve autonomous and foresighted scheduling mode in more complex scheduling scenarios.However,intelligent scheduling can be difficult to achieve when virtual data is not effectively mapped and integrated with real physical workshops.To explore intelligent scheduling mode,this paper studies the flexible job shop scheduling problem based on digital twin.The contents are as follows:(1)A job shop scheduling mode based on digital twin is proposed.In order to make full use of production data to guide job shop scheduling,and to achieve the scheduling mode with features such as autonomy and rapid response,a job shop scheduling mode based on digital twin is proposed.The overall framework of the scheduling mode based on digital twin and the intelligent scheduling decision service is designed with the characteristics of the digital twin.The operating mechanism of the main components of the intelligent decision service are introduced in detail.(2)A hybrid competitive swarm optimizer is proposed to generate initial scheduling scheme.A hybrid competitive swarm optimizer is proposed to make up for the shortcomings including high computational complexity and difficulty in solving flexible job shop scheduling of the existing methods.In order to enhance both the global and local search ability,the POX crossover(Precedence Operation Crossover),ring topology and neighborhood search are applied to the update of the winners.Lastly,through case testing and comparison with other algorithms,it is proved that the hybrid competitive swarm optimizer has high solution stability to the flexible job shop scheduling problem.(3)A dynamic scheduling decision method powered by digital twin is designed.In order to make the intelligent scheduling decision service meet the needs of actual scheduling,a dynamic scheduling decision method driven by digital twin is designed.Firstly,a rescheduling strategy based on real-time status data is designed,and implicit disturbances are converted into random events by using Siamese Network to achieve more precise rescheduling strategy;Then,through the Pseudo-Siamese CNN,the mapping relationship between the process and machine state,which is then applied to guide rescheduling decisions,is obtained.Finally,the validity of the dynamic scheduling decision method driven by digital twin is verified through simulation analysis.(4)A scheduling prototype system is built.Based on the above research results,a scheduling prototype system is developed based on Java Script.And the system includes machine management,workpiece management,algorithm management,log management,user management,task management,initial scheduling,dynamic scheduling and status monitoring.