Research On Resource Scheduling Model Of Industrial Internet Of Things Based On Osmotic Computing

Due to the rapid development of the Internet of Things,more and more industrial equipment are participating in the IIo T.The increasing scale of Io T has significantly increased the data flow of Io T.Thanks to deployment in the cloud far away from users,there are problems such as high latency,low bandwidth threats.For the limited computing power and storage capacity of edge computing,there is an urgent need for a strategy to skillfully combine the advantages of cloud computing and edge computing.Therefore,the cloud-edge collaborative computing model has gradually become one of the current research hotspots in the IIo T field.In the traditional cloud computing and edge computing frameworks,resource scheduling problems have received extensive attention,and related theories,technologies and algorithms have developed more maturely,and the research on resource scheduling problems in cloud-edge collaborative computing frameworks is still in its infancy stage.The thesis aims to address the problems of IIo T massive heterogeneous resource scheduling and processing specific needs under the cloud-edge collaborative computing architecture,and proposes a resource scheduling method that can meet low latency,low energy consumption and high efficiency.The main research contents and contributions of this thesis are twofold:(1)An IIoT-oriented cloud-edge collaborative resource scheduling model based on osmotic computing is proposed.Firstly,according to the computing characteristics of cloud computing and edge computing,a collaborative cloud-edge resource osmotic resource scheduling model is proposed,and Qo S is introduced to achieve adaptation to multiple indicators such as response time,throughput,and energy consumption.The adaptation process adopts the Pareto effective configuration idea.Then,a semi-permeable membrane osmotic management mechanism based on deep reinforcement learning is proposed which uses deep reinforcement learning algorithms to act as a scheduling manager between cloud-edge resources,so that service resources can achieve optimal configuration of service resources and efficient use of cloud-edge resources under the constraints of multi-dimensional Qo S Pareto effective configuration.Finally,the proposed cloud-edge resource scheduling algorithms are compared and analyzed through simulation experiments.(2)An IIoT-oriented edge-edge collaborative resource scheduling model based on osmotic computing is proposed.Firstly,the service priority is divided according to the delay index of the service resource,and then the computing power,processing time,and energy consumption factors are comprehensively considered to construct the evaluation index of multi-dimensional Qo S Pareto effective configuration.Then,a method based on improved deep reinforcement learning is proposed.The semi-permeable membrane management mechanism of the algorithm improves the convergence speed of the deep reinforcement learning algorithm by introducing a LSTM network which helps DQN learn the best strategy faster,and realizes the optimal configuration of service resources and edge servers.Finally,the proposed edge-edge collaborative resource scheduling algorithms are compared and analyzed through simulation experiments.