Research On Fog Resource Scheduling Technology For Intelligent Manufacturing Services

Smart manufacturing opens the era of fog computing analysis,where fog computing drives storage,transmission,and computing between terminal devices and storage data and cloud computing.With the development of smart manufacturing,fog computing,and Internet of Things technologies,there have been more and more mart manufacturing services,such as industrial big data analysis,intelligent maintenance and management,flexible factories,and other smart applications.This thesis analyzes the characteristics and requirements of intelligent manufacturing services,and proposes a fog computing architecture for smart manufacturing services.Through the design of smart manufacturing service application classification and communication interaction process,the architecture can reasonably provide smart manufacturing services.In addition,network element entities such as fog clusters,fog element nodes,and fog management nodes were introduced,and a clustering algorithm for fog resource management was proposed to facilitate the effective cooperation of fog layer resources.Based on the proposed fog computing architecture,this project aims at the characteristics of equipment computing capacity,delay requirements,and communication support capabilities in smart manufacturing,and proposes a fog computing resource scheduling model based on constraint tasks,with time delay and communication load as the optimization goals.And then,two fog resource scheduling algorithms are proposed,namely Pri-Min algorithm and ADGRS(Adaptive Doublefitness Genetic Resource Scheduling)algorithm.The Pri-Min algorithm uses a non-heuristic algorithm and draws on the idea of the classic Min-Min algorithm to calculate the constraint relationship between tasks as the task's priority,and then based on the priority for resource scheduling.This algorithm takes the weighted benefits of delay and communication load as the optimization goal,and can optimize the delay and communication load in different weights.The Pri-Min algorithm produces a local optimal solution,which is not necessarily optimal in the global scope.So this thesis proposes the second scheduling algorithm ADGRS.This algorithm belongs to the heuristic algorithm and adopts an adaptive genetic algorithm with double fitness.This algorithm will find the approximate optimal solution in the global space,and also can optimize the delay and communication load in different weights.The algorithm simulation and result analysis show that Pri-Min algorithm and ADGRS algorithm can both take into account the delay performance of the task and the overhead of communication resources,which is beneficial to the effective cooperation and full use of fog resources.Compared with greedy algorithm and OLB(Opportunistic Load Balancing)algorithm,Pri-Min algorithm has better comprehensive performance.Compared with the Con-Min(Constraints Min)algorithm,the ADGRS algorithm has better comprehensive performance.The Pri-Min algorithm schedules faster than the ADGRS algorithm,but the comprehensive scheduling results are weak.Therefore,these two algorithms have their own advantages and disadvantages,which can be applied to different demands of the application.