Research On Task Dynamic Assignment Method In Cloud-Fog Collaborative Environment

With the rapid development of Internet of Things technology,the data generated by terminal devices has shown explosive growth,which has brought enormous pressure to cloud computing.The emergence of fog computing has relieved the pressure of cloud computing,and at the same time made up for the shortage of terminal equipment in terms of storage resources and computing power.How to utilize all computing resources to improve user experience quality is one of the important challenges in fog computing.At the same time,the most essential feature of terminal equipment is mobility,and the coverage of fog nodes has certain limitations.Every movement of a device may cause changes in the group of fog nodes it is connected to.Therefore,how to It is a problem worthy of research to allocate the tasks of the terminal equipment to the corresponding computing nodes reasonably and efficiently to ensure the user’s quality of experience.In view of the above two problems,the main work of this paper is as follows:(1)The task allocation method with the optimization goal of delay and energy consumption in the scenario of single fog node and multi-terminal equipment is studied.Firstly,the task allocation problem based on joint optimization of delay and energy consumption is established as a 0-1 integer programming model.Secondly,a binary beetle antennae search algorithm(Binary Beetle Antennae Search Algorithm,BSBAS)is proposed,and the Sigmoid function is added on the basis of the beetle antennae search algorithm,and the updated position of the beetle antennae is mapped to 0 or 1 with 0.5 as the boundary.Finally,simulation experiments are carried out with polling algorithm,random algorithm,all offloading methods and all execution methods on terminal devices as the baseline algorithms.The simulation results show that whether the number of terminal devices changes or the computing power of fog nodes and terminal devices changes,the proposed algorithm has the best performance in the task dynamic allocation problem.At the same time,the experiment also verified the impact of the weight on the delay and energy consumption.If the weight is biased towards the delay,the algorithm will optimize the delay more than the energy consumption,and vice versa.(2)The task allocation method with the average delay of the task as the optimization goal in the scene of multi-fog nodes and multi-terminal devices is studied.Firstly,the advantages of software-defined network(Software Defined Networking,SDN)are applied to the cloudfog-device architecture,and the dynamic allocation of tasks is made more reasonable and efficient by providing a solution for centralized control of information such as terminal devices,fog nodes,and cloud server resources.Secondly,a task prioritization method based on the K-means algorithm is proposed.First,the tasks of the terminal equipment are prioritized,and then a task allocation method based on the improved quantum pigeon-inspired optimization algorithm(Improved Quantum Pigeon-inspired Optimization,DQPIO)is proposed.Assign tasks to prioritized task sets.Finally,the proposed method is simulated in different scenarios,and the results show that,compared with the polling algorithm,beetle whisker search algorithm,pigeon group optimization algorithm and random generation algorithm,the proposed method is more effective than the terminal equipment.The number,data volume of tasks,and the number of fog nodes have all achieved the best performance.(3)The task allocation problem with the optimization goal of minimizing the delay is studied in the scenario of considering the mobility of the terminal equipment.Firstly,the tasks generated by the terminal equipment are constructed as a directed acyclic graph,and an optimization model of workflow task allocation with the goal of minimizing time delay is introduced.Secondly,in order to maintain the continuity of the task during the mobile process of the terminal equipment,when the terminal equipment moves to the sensitive area,the task migration decision-making mechanism is designed according to the current computing resources of all computing nodes and the migration cost of the task.A task allocation method for the Deep Q Network(Deep Q Network,DQN)algorithm.Finally,through simulation experiments,the effectiveness of the proposed method is verified in terms of the number of tasks and the amount of data of the tasks using the Q-learning algorithm as the baseline algorithm.