| Fog computing provides a distributed,close of calculation,storage and network resources for tasks related to intelligent manufacturing.As an extension and promotion of the cloud computing,fog computing has the characteristics such as low latency,low energy consumption,which is an emerging technology to realize the mass manufacturing data real-time analysis and complex task real-time processing.Intelligent terminal equipment,fog node,and cloud server provide different granularity computing resources for intelligent manufacturing,effectively using these resources to improve the real-time,energy efficiency and reliability performance of intelligent manufacturing related tasks,which is the main purpose of implementing fog computing.Compared with the traditional manufacturing,intelligent manufacturing features such as intelligence,flexibility making tasks with heterogeneity.The high-performance processing of heterogeneous task bring challenges,puts forward the higher request of its key technology.This paper improves the real-time,energy efficiency and reliability of the three performance indexes of intelligent manufacturing related tasks,the key technologies such as fog nodes deployment,computing mode selection and fog computing task scheduling have researched.The main research contents and innovation points are as follows:(1)The effective analysis of the fog computing architecture in the intelligent factory is the prerequisite to realize task real-time and efficient processing.Combined with fog computing architectures of intelligent factory,three computing modes of fog computing architectures,namely local computing mode,cloud computing mode,fog computing mode is analyzed and the system modeling.Task processing of different computing modes has studied,the latency,power consumption and reliability mathematical models of each computing mode have established.The simulation results show that three kinds of computing mode in latency,power consumption and reliability have different performance results.(2)The effective deployment of fog nodes is the foundation of building fog computing platform.Intelligent manufacturing system is analyzed in time and space characteristics,this paper proposes a fog nodes deployment strategy based on temporal and spatial characteristics.The objective function of fog node deployment is established,which is to minimize the response time and realizing load balancing,the discrete differential evolution algorithm is adopted to obtain the optimal solution.The simulation results show that the proposed fog node deployment strategy has achieved the goals on reduce response time and load balance.(3)In the fog computing environment fused by local computing mode,fog computing mode and cloud computing mode,making the best computing mode options for the tasks is the key to ensure that a high real time,low energy consumption,high reliable task execution.This paper proposes an adaptive computing mode selection strategy based on task priority,a computing mode selection module is designed in the fog server.The genetic algorithm is used to solve optimization computing mode selection,through the strategy can set the optimal calculation model for task options.In order to guarantee and improve the real-time and reliability of delay sensitive tasks under fog computing mode,an adjustment mechanism of task execution order is proposed based on task priority.Based on task priority adjustment task queue,so as to realize the high priority task first get fog computing services.In addition,the merge sort algorithm is adopted to accomplish the task queue fast adjustment,the algorithm solved the problem of long time and low accuracy on the traditional sort algorithm.The simulation results show that the proposed computing mode selection strategy has obvious advantages on real time,energy efficiency and reliability.(4)Aiming at the problem on fog computing task exist high delay and high energy consumption,considering the heterogeneous features of fog computing tasks and fog nodes,a fog computing task scheduling strategy based on level mapping hybrid heuristic algorithm is proposed.According to the requirements of the task performance computing task priority,realizing fog computing tasks hierarchy based on priority.According to the attribute evaluation of fog node computing executive force,completed the fog nodes hierarchy based on executive force.Task sets with the same magnitude and fog node set to establish a mapping relationship,reduce the search space of optimized algorithm,improves the execution efficiency of task scheduling.A hybrid heuristic algorithm is adopted for task scheduling optimal solution.The simulation results show that through the proposed scheduling strategy task latency,power consumption and reliability performances are effectively improved.On the basis of personalized production candy packaging intelligent production line designed by laboratory,a fog computing implementation verification platform in intelligent manufacturing is set up,strategies and algorithms of this paper are verified on the platform.Based on task latency,power consumption and reliability of the performance test and analysis,validate the consistency of experimental results and simulation results.The key technology research of fog computing providing theoretical reference for fog computing application in the field of intelligent manufacturing. |
