Resource Allocation Method Of Edge Computing Networks For Intelligent Manufacturing

Edge-cloud network architecture has been widely adopted in industrial production scenarios.Industrial enterprises closely connect factories,warehouses,equipment and customers through computing and communication-integrated platforms to share pluralistic resources in the industrial production processes.Obviously,the computing power of intelligent production and transportation equipment has made great progress,and they have a certain ability of original data collection,screening and preprocessing,which makes it possible to facilitate decentralized technology development and smart intelligent algorithm design at the edge level.Especially for the typical requirement such as low-delay production beat and scarce resource competition game,the technologies and strategies such as the task segmentation,resource slicing,collaborative optimization and Nash negotiation are able to realize the parallel processing of subtasks and the optimal allocation of industrial computing resources in some production scenarios.Firstly,in recent years,the customized production and collaborative computing services based on mobile edge computing have shown explosive growth.Similarly,driven by 5G and other broadband access technologies,the proportion of single large volume services in the overall services is becoming higher and higher,and the requirements for processing speed and complexity are becoming more and more obvious.Therefore,orienting to the problem of increasing delay caused by service cutting and result aggregation of distributed computing in the edge micro data center network,this thesis proposes a mobile edge computing(MEC)collaborative scheduling algorithm based on a hierarchical data architecture and subtask differential aggregation.By dynamically changing the order of task data aggregation in the time-space dimension,the edge micro data center is divided into different clusters for data aggregation,which effectively avoids the congestion caused by the simultaneous aggregation of multiple tasks to a single edge data center.Further,this thesis designs the path selection and timing strategy of subtasks in the cluster,and introduces the idea of directional penetration in chemistry.The processing time difference and fine-grained spectral fragments of several subtasks are regarded as osmotic pressure,so that the processed subtasks are entered in aggregation stage in advance.Finally,simulation tests are carried out in different topologies to verify the advantages of the proposed algorithm in terms of average service completion time,aggregation efficiency and spectrum utilization.Subsequently,in the MEC scenarios such as intelligent factory and intelligent Wharf in industrial Io T,based on the server customer bilateral mutual selection strategy of multiple resource indicators,this thesis provides an edge computing service equilibrium selection algorithm for intelligent production equipment such as industrial automated guided vehicle(AGV)and manipulator.Firstly,based on the service popularity and Ziff’s law,determine the process and main service type of each remote unit(RU)in the network.At the same time,divide all Ru of the network based on the service heat similarity between Ru to obtain the working area slice.Then,the AGV side first sends a forward service request to the Ru side based on the task content and urgency,and the Ru side then reversely selects the AGV from the perspective of the optimal allocation of resources in the whole network.If AGV receives that all feedback information in the alternative Ru list is insufficient resources,it will send the service request to the cloud.This thesis coordinates multiple elements such as service content matching,user delay limit,computing and communication resources,and tries to solve the problem of dynamic allocation of limited resources in edge cloud computing network to AGV service requests in the way of balanced matching under scenarios such as factories and ports.At the same time,this thesis adopts the strategy of positive and negative two-way selection of agv-ru to solve the game problem of user operator in service experience and benefit maximization.Through simulation experiments,the radio resources,optical fiber link resources,spectrum resources,computing resources and the average service delay of the whole network under the above scenarios are compared and analyzed to verify the performance advantages of the proposed algorithm.This thesis focuses on the sub-task aggregation problem in the distributed computing of general production tasks and the accurate resource matching problem of the edge industrial network,trying to feed back the large-scale,low-delay and high-precision production instructions to the equipment on the production side,so as to not only promote the productionoriented enterprises to effectively improve the production sequence and process in intelligent manufacturing and warehousing,but also provide algorithm support on the formulation of production decisions in the future.