Research On 3D Printing Resource Task Matching Method In Cloud Manufacturing Environment

As a new high-end rapid manufacturing technology,3D printing technology has brought new opportunities for the upgrading of China’s manufacturing industry.With the vigorous development of 3D printing technology,3D printing has been applied in more and more industries,bringing higher benefits to countries and enterprises,and bringing more convenience and creativity to individuals.The uneven distribution of3 D printing resources and the constraints of manufacturing service capabilities have led to the coexistence of idle resource capacity in some regions and insufficient resource capacity in some regions,exacerbating the contradiction between supply and demand between resource demanders and resource providers.Cloud manufacturing is a product of the deep integration of modern information technology and manufacturing industry.Its core idea is to aggregate scattered and idle manufacturing resources onto a cloud platform to facilitate resource demanders to obtain resources on demand through network terminals.This article combines the two,using a cloud manufacturing platform to aggregate distributed 3D printing resources,in order to improve the efficiency of 3D printing and the utilization of 3D printing resources.(1)A 3D printing resource task matching model considering both parties’ preferences is proposed.Firstly,a basic constraint model is constructed based on the basic attributes of 3D printing resources and 3D printing tasks;Then analyze the preferences of both parties,use the entropy weight method to calculate the weight of preferences,and construct a preference income function for both parties.By adding the relationship between preferences to adjust parameters,the preference income function is converted into a comprehensive income function,which is solved using particle swarm optimization.The experimental results show that the 3D printing resource task matching model considering both parties’ preferences is more stable than the matching scheme based on unilateral decision-making,and adding stability constraints can further improve the stability of the matching pair solution set.(2)A task matching model for 3D printing resources based on deep reinforcement learning is proposed.In the cloud manufacturing environment,the 3D printing resource task matching environment is often dynamic and complex.In the face of unknown environments,deep reinforcement learning will rely on continuous trial and error learning to optimize decision-making options,introducing deep reinforcement learning into the 3D printing resource task matching environment,providing new ideas for solving the 3D printing resource task matching problem.Firstly,3D printing resources and task attributes are analyzed,and a Qo S optimization function is constructed;Then,a 3D printing resource task matching model based on deep reinforcement learning is constructed.Finally,a simulation experiment is conducted using the deep Q network algorithm,and the experimental results are compared with the other two heuristic algorithms.The experimental results show that the deep Q network algorithm can effectively solve the 3D printing resource task matching problem,and is superior to the other two heuristic algorithms.(3)A web-based 3D printing resource task matching system was designed and developed.The system intelligently identifies the number of tasks submitted by the task party and selects different matching models to increase the user experience.When a task party submits a single task,the system will select a 3D printing resource task matching model that takes into account both parties’ preferences;When the task party submits multiple tasks,the system will select a 3D printing resource task matching model based on deep reinforcement learning.Firstly,different functions are assigned to the three roles of resource party,task party,and administrator in the system,and then the system database is designed.Finally,functions such as resource upload,resource management,task upload,and intelligent matching are implemented.