Research On Edge Computing Node Deployment And Energy Saving Technology For Intelligent Factory

In the era of Industry 4.0,the explosive increase in the number of Internet of Things(IoT)devices in factories,the rapid changes in network architecture,and the demand for unmanned business have promoted the transformation and upgrading of intelligent factories around the world.However,the network architecture of the traditional intelligent factory is usually cloud-centric,and the data and information from the equipment are processed in the cloud.This centralized management method cannot support the real-time computation of large-scale IoT devices in the IoT paradigm but will lead to heavy load on the cloud,which greatly increases the response delay of service requests.To well support the construction of the IoT paradigm in the intelligent factory,edge computing,as an edge extension of cloud computing model,is applied to the construction of the intelligent factory network architecture.As Cloud-Radio Access Network(C-RAN)can meet the requirements of IoT for network flexibility and scalability,this thesis proposes an intelligent factory edge computing architecture based on C-RAN.In this thesis,the traditional optimization algorithm and heuristic algorithm are combined to solve the edge computing deployment problem in intelligent factory.Based on the C-RAN intelligent factory edge computing architecture,this work considers the offloading mode of a single sensor device to offload computing tasks to multiple edge access devices and constructs a non-linear programming mathematical model for jointly optimizing the deployment cost of equipment and the data offloading delay of smart services.In this case,a hybrid algorithm based on the immune algorithm and the Lagrange multiplier method is proposed to solve this non-linear joint optimization problem.Compared with the offloading mode in which a single sensor device offloads the computing task to a single edge access device,the proposed offloading mode in this thesis can effectively reduce the deployment cost and data offloading delay when adopting a hybrid algorithm based on the immune algorithm and the Lagrange multiplier method.Aiming at the problem of offloading energy consumption and delay optimization of decision-making task flow for mobile devices in intelligent factories,this thesis adopts a combination of multiple heuristic algorithms to solve it.Based on C-RAN intelligent factory edge computing architecture,this thesis proposes five different offloading modes:local offloading,edge offloading,cloud center offloading,local-edge collaborative offloading,and local-edge-cloud collaborative offloading.After mathematically modeling the energy consumption and delay joint optimization problem,this thesis proposes an algorithm based on genetic algorithm and particle swarm optimization algorithm to deal with this problem.The experimental result proves that the proposed algorithm can obtain the minimum task flow offloading energy consumption and delay.