Task Scheduling And Resource Allocation In Mobile Edge Computing

Currently,computationally intensive applications such as interactive gaming,augmented reality,facial recognition and image/video processing are becoming popular on mobile devices.The rapid growth of mobile applications places stringent demands on cloud infrastructure and wireless access networks,such as user experience continuity,ultra-low latency and high reliability.These requirements drive the need for highly localized services at the edge of the network near end users.In view of this,Mobile-Edge Computing(MEC)as an emerging computation paradigm is considered as a key technology to solve these problems.Among the main research directions of mobile edge computing,computational offloading has received great attention.Compared to the cloud,the MEC server usually has finite resources.As a consequence,MEC server would not be able to meet all users' computation requirements.Therefore,the joint optimization of offloading decisions and resource allocation are important research issues in MEC systems to improve the network performance.Although there are many studies on computation offloading for MEC systems,these works are unsuitable for the problem of joint access and backhaul bandwidth allocation.In practice,such wireless backhaul constraints exist in wireless MEC heterogeneous networks and they affect the offloading decisions.When a mobile device decides to offload its task to a mobile edge server,the nearest mobile edge server is not always the best offloading target.Despite the short distance,if there is congestion in the network and the overhead of the mobile edge server is high,it will not be able to meet the requirements of the task.It is important to consider communication and processing time and processing requirements when offloading the appropriate mobile edge server.In response to the problems in MEC,the main contributions of this paper are as follows:(1)By combining the advantages of communication technology and unmanned aerial vehicles(UAV),we use UAV as mobile base stations.UAV can offload tasks to mobile edge servers that can provide computing resources.The mobile edge server can be a cellular base station,a Wi-Fi access point,or the like.When a mission occurs in a specific area,one or more drones need to move locations to acquire data and process it.If data processing is too tedious to process locally,UAV can work with mobile edge servers.Based on the Hungarian algorithm,which is one of the matching algorithms,this paper proposes an optimal mobile device-UAV-edge server matching algorithm,we consider the computing capacity of edge server and the task queue comprehensively,choosing the best edge server for mobile users,and by using the UAV as a mobile wireless network access point,the task is executed with minimal computing overhead.and verified by simulation experiments which can minimize energy consumption and processing time.(2)In order to minimize energy consumption while meeting task completion time constraints,considering channel resource allocation for task data upload and backhual,this paper provides an energy-efficient and efficient dynamic offloading and resource allocation strategy to reduce computational overhead.Consider placing a mobile edge server in a cell,we first proposed optimization problems for mobile users' offload strategy,bandwidth allocation,and computing resource allocation.In order to solve the optimization problem proposed,we decomposes it into two sub-problems to be solved separately,and then iterates to obtain the optimal solution.Dynamic Voltage and Frequency Scaling(DVFS)is widely applied in most current research,multi-core processors bring new performance improvements based on DVFS.We combine high-performance and high-power large CPU cores with small CPU cores that are low-performance but energy-efficient.Therefore,we can reduce the task execution overhead by rationally distributing the execution of tasks on different cores.