Mobile Edge Computing Fine-grained Resource Scheduling And Task Offloading

In the research field of mobile edge computing,there are more research on scenarios where the system has multiple servers and allows simultaneous task offloading,and less research on fine-grained resources of edge node servers.The former is often difficult to be adopted for MEC systems with limited resources in real life.Therefore,this thesis focuses on the task offloading scheduling problem of fine-grained resources in the system under resource constrained scenarios.First,to solve the problem of system delay deterioration caused by the coupling of multi-task transmission and execution process in mobile edge computing,this thesis proposes a dual queue task offloading method.On the basis of flexible flow shop scheduling and genetic algorithm optimization to solve the minimum delay energy trade-off sum,the delay constraints of each task are considered,Divide the task offloading queue into a pre queue that does not meet latency constraints and a post queue composed of other tasks.By utilizing genetic algorithms to optimize the pre queue and post queue separately,a reasonable task offloading order is ultimately obtained,ensuring that tasks that do not meet delay constraints can be executed before other tasks and return results to user devices earlier.A comparative analysis was conducted on the latency advantages of this method for task offloading,the variation of overall queue latency with the increase of fine-grained server resources,and the trade-off relationship between latency and energy consumption.The performance of the algorithm was evaluated,and the results showed that the algorithm used in this thesis solved the latency and energy consumption trade-off better than the comparison algorithm.Secondly,in order to solve the problem that users cannot obtain server resources for a long time when multiple users offload tasks at the same time in mobile edge computing,this thesis designs the server task scheduling algorithm based on the classic process scheduling method of operating system and the characteristics of task offloading in MEC system The performance performance of the system using multiple server task scheduling algorithms,allowing multiple tasks to be transmitted simultaneously and executed simultaneously on the server.Considering the fine-grained resources of the server,the impact of each scheduling algorithm on system latency was compared and analyzed under different server processor counts and processor execution frequencies.Finally,the trade-off between system latency and system energy consumption was simulated and analyzed,revealing that task offloading scheduling is more critical when radio and computing resources are relatively balanced,that is,when both parties do not dominate the other,The task uninstallation process is smoother.Finally,this thesis designs and implements a task offloading solution system for edge computing to solve the problem of system delay deterioration in resource constrained scenarios.The system has the function of adding and deleting devices and nodes to facilitate users to build edge computing scenarios;With the delay solving function of task queues,it is convenient for users to obtain the delay results of input task queues;Equipped with task offloading queue recommendation function,it can minimize latency and recommend offloading sequence for tasks selected by users,making it convenient for users to reduce latency overhead.