At present,cloud manufacturing,as a result of the integration of advanced technologies such as cloud computing,the Internet of Things,and artificial intelligence,has had a profound impact on the manufacturing industry.Cloud manufacturing has not only attracted widespread attention from the academic community but also governments and industry have invested a lot of money to research it.The results,its theoretical and methodological studies have important academic and applied value.Under the cloud manufacturing model,manufacturing resources can not only achieve greater sharing but also reduce manufacturing costs while improving the utilization of manufacturing resources.This paper studies the optimization problem of cloud manufacturing service composition and the scheduling problem of dynamic high-dimensional manufacturing resources under multi-objective conditions in cloud manufacturing.The main results of creative research are:(1)Aiming at the problems of fragmented manufacturing services and single service functionality,this article studies the cloud manufacturing service composition.In view of the multiple optimization target problems that arise in the actual situation of the cloud manufacturing service composition,a multi-objective cloud manufacturing service composition optimization model was established,taking into account four indicators of cost,time,service quality,and energy consumption,and an evaluation loss function tuning was proposed.rule.Through problem analysis,a fast adaptive grid multiobjective service composition algorithm based on variable neighborhood search is proposed.The algorithm does not need to manually set the size of the neighborhood search range,and can automatically use the set grid to delete a large number of dominated individuals.The simulation results show that the algorithm effectively improves the problem of uneven Pareto solution set when optimizing multiple targets,and improves the speed and accuracy of cloud manufacturing service composition.(2)Aiming at the problem of dynamic resource distribution and difficult optimization in the manufacturing process,this paper studies the dynamic resource scheduling of cloud manufacturing.In view of the problem of improper task scheduling and waste of resources in the actual situation of cloud manufacturing resource scheduling,a dynamic production resource scheduling optimization model in the cloud manufacturing environment was established,and manufacturing fuzzy evaluation rules were designed.Based on the problem analysis and the reinforcement learning algorithm,a deep reinforcement learning resource scheduling algorithm based on tabu search is proposed.The algorithm controls the sampling space through a blacklist management mechanism,and makes the benign and malignant actions easier to distinguish through the definition of adaptive baseline values. |