Computation Offloading And Resource Allocation Methods In Power-Constrained Edge Computing Systems

By utilizing the cloudlets deployed at the edge of networks,mobile edge com-puting(MEC)provides ample computing and storage services for users to make data generated locally and consumed locally.Compared with the conventional mobile cloud computing(MCC),which relies on the remote data centers to provide services,MEC has the following advantages:First,MEC servers are deployed nearby users,which greatly shortens the transmission delay caused by the long distances between users and the data centers in MCC systems.Therefore,real-time application services can be deployed in MEC servers.Secondly,the distributed deployment of MEC servers avoids the heavy burden of transmitting the massive information generated locally to the data centers through the existing core network.Third,MEC servers are closer to users and can gather real-time information of the end users,such as behavior,location,and environmental information,thereby providing better service to users.Therefore,MEC can provide a low-latency computation offloading service for users,assist in the collection of massive data at the edge of networks,and optimize the network flow based on the environmental information.MEC is one of the key technologies to solve the contradiction between distributed collection and centralized processing of information and realize the vision of the Internet of Everything.Based on the above reasons,MEC has attracted widespread attention in academia and industry in recent years.This dissertation focuses on the computation offloading and resource allocation methods in power-constrained MEC systems.In addition,the MEC-assisted Peer-to-Peer(P2P)communication between mobiles is analyzed.On the one hand,the task-based computation offloading methods in the existing MEC theory can achieve a good trade-off between the energy consumption and the processing delay but do not pay enough attention to the limited available power of users.On the other hand,the task-based computation offloading methods require the network controller to fully know the task details,thus cannot deal with the low-latency tasks in the IoT networks.There-fore,it is necessary to quantify the available computation resources of users to achieve resource reservation for low-latency services.In addition to increasing the computing capacity of edge devices,MEC can also be utilized to help communication.This work consists of three parts,which are respectively studied in single-user MEC systems,multi-user MEC systems and MEC-assisted computation relaying systems.Each part contains several research points.The main contributions are summarized as follows.(1)First,the APC which describes the computing ability and speed of a served IoT device is defined.Then its expression is derived by analyzing the relationship between task partitioning and resource allocation.Based on this expression,the power allocation solution for the single-user MEC system with a single subcarrier is studied and the factors that affect the APC improvement are considered.The analysis results show that rich computing resource is provided to users by MEC,which greatly enhances the information processing capability of the network edge.(2)For the multiuser MEC system,an optimization problem of APC with a gen-eral utility function is formulated and several fundamental criteria for resource allocation are derived.By leveraging these criteria,a binary-search water-filling algorithm is proposed to solve the power allocation between local CPU and multiple subcarriers,and a suboptimal algorithm is proposed to assign the subcarriers among users.For mobile devices,especially IoT devices,high energy efficiency can greatly extend their standby time.Therefore,the power and subcarrier allocation algorithms are further derived by maximizing the computing energy efficiency in multi-user MEC systems.The APC-based resource allocation strategy is more suitable for the deployment of real-time services and the reservation of computing resources.(3)An MEC-assisted computation relaying scheme is proposed to enhance the throughput of uncompressed data for mobile peer-to-peer(P2P)communications.The original data is assumed to be transmitted from one mobile device to another through a relay node equipped with MEC servers.During the transmission,the data compression rate can be changed by either the mobile devices or the relay node.The transmitting power and compression rates are adapted to the computing and communication resources in the system.To assess the system performance,a cost function defining the tradeoff between energy consumption and latency time is used.By minimizing the cost function,the optimal transmission and compression strategy is derived and analyzed with practical constraints on computation rates,MEC server,energy consumption,and latency.Compared with the conventional scenario without MEC,the proposed model can break the bottleneck of P2P communications by decoupling the data compression rate in the two relaying hops.