Research On Cloud Manufacturing Service Recommendation System Based On Graph Neural Network

Intelligent manufacturing is the core of Made in China 2025 and the main direction of development of China’s manufacturing industry.Intelligent service is one of the core contents of intelligent manufacturing.As a manufacturing resource service platform,cloud manufacturing is gradually becoming an effective carrier of intelligent manufacturing services.With the rapid increase of resources on cloud manufacturing service platforms,it is difficult for users to quickly and accurately find the required manufacturing service resources from the platform.Therefore,recommendation systems have become a key module of cloud manufacturing service platforms.This project proposes a cloud manufacturing service resource recommendation method based on graph neural networks to address the issue of existing cloud manufacturing service recommendation methods not utilizing high-dimensional feature information of resources to improve recommendation accuracy.A cloud manufacturing service recommendation prototype system based on this method has been developed.The main research contents are as follows:(1)Analysis of problems in cloud manufacturing service resource recommendation system.The study firstly analyzes the recommendation problem of cloud manufacturing service resources systematically,including the analysis of user requirements of cloud manufacturing service platform and the description of information overload problem of cloud manufacturing service platform.At the same time,the research focuses on the detailed analysis of the graph neural network-based manufacturing service resource recommendation method and recommendation process.(2)Establishing a recommendation model for cloud manufacturing services based on graph neural networks.The study uses different similarity calculation methods to calculate the similarity between manufacturing service resources and establishes a cloud manufacturing service resource graph dataset.Based on this,a graph neural network-based link prediction model for cloud manufacturing service resources is constructed,and the model is trained using the established graph dataset.Finally,the experimental results of the model are analyzed and evaluated.(3)Analysis and design of cloud manufacturing service recommendation system.Based on the established cloud manufacturing service recommendation model,the cloud manufacturing service recommendation prototype system has carried out requirements analysis,functional module division,use case analysis and timing diagram analysis,as well as system architecture design,class design,database design,etc.(4)Functional implementation and test evaluation of cloud manufacturing service recommendation system.Based on the system analysis and design,the key functional modules such as resource release,demand release and personalized recommendation of manufacturing services in the cloud manufacturing service recommendation prototype system were implemented,and functional testing and effect evaluation were carried out.The research results indicate that the proposed manufacturing service recommendation method can achieve good recommendation results and accurately recommend suitable manufacturing service resources to users on the platform.