Online Optimization Scheduling Of 3D Printing Manufacturing System In Multi-task Printing Mode

With the rapid rise of 3D printing technology and the rapid development of Internet technology,3D printing manufacturing systems come into being.The user releases the print job via the Internet.The 3D printing devices access the manufacturing network through the Internet of Things,and receive printing tasks issued by users through the Internet as service providers.This kind of networked and personalized manufacturing is becoming a new development direction of the manufacturing industry.Therefore,the study of the optimization scheduling problem of 3D printing manufacturing system has important practical significance for improving the utilization efficiency of 3D printing equipment.3D printing manufacturing system under the multi-task printing mode belongs to a special online optimization problem of single machine batch considering the workpiece size.Firstly,the rectangular layout algorithm is used to solve the problem of the combination layout of different task models in multi-task printing mode.By analyzing the printing process of 3D printers using FDM forming technology,the calculation formulas such as printing time and energy loss in the printing process are given.Then a semi-Markov decision process(SMDP)is established.And the policy iteration algorithm is used to solve the system optimal strategy.At the same time,this dissertation gives a model-free Q-learning algorithm based on simulated annealing to optimize the system.By analyzing the simulation results,it is verified that the optimization algorithm improves the utilization of the system's printing space.The printing parameters have an important influence on the printing process and the quality of the printed task model.When the printing parameters are fixed,the 3D printing device cannot be fully utilized.Therefore,the online optimization scheduling problem of 3D printing manufacturing system in the multi-task printing mode with variable printing parameters is studied in this dissertation.The task combination and parameter layer height are viewed as the joint control variables of the system.And a semi-Markov decision process(SMDP)is established.Policy iteration algorithm and Q-learning algorithm based on simulated annealing are adopted to optimize the system.By comparing the performance of the system under variable and fixed parameters layer,it shows that the system with variable parameter layer has higher flexibility.