Research On Resource Allocation Based On Collaborative Edge Computing

With the widespread application of Internet of Things technology in life worldwide,smart devices are playing an important role in scenarios such as smart cities,autonomous driving,and Industrial Internet.As the amount of data generated by smart devices in-creases exponentially,there are more and more tasks with low latency,high storage,and complex computing requirements.The communication efficiency between the smart net-work device and the cloud computing center is low and the communication cost is very high.The emergence of Edge Computing(EC)technology aims to extend cloud comput-ing functions to the edge of the network close to user terminals,bringing users short-range computing and storage capabilities.By offloading the task request of the user terminal to the Edge Computing Server(ECS)for processing,the task processing delay and energy consumption of the user terminal can be significantly reduced.Because EC technology is a distributed architecture,each ECS runs independently.When user-generated tasks have complex requirements for computing and storage re-sources,a single ECS is not enough to provide all the resources it needs.Collaborative EC can dynamically allocate resources by combining user terminals,ECS,and cloud servers to provide services under the constraints of user delay and energy consumption.However,the current research content based on collaborative EC scenarios is less and most of them are modeled as non-convex optimization problems.How to design effective collaboration strategies and corresponding algorithms is still very challenging.This thesis studies the problems of computing offloading and resource allocation,request receiving and content caching in collaborative EC scenarios.The main work and innovations are as follows:First,given the high processing delay and unreasonable resource allocation in cur-rent EC strategies,we propose a joint optimization strategy for computing offloading and resource allocation based on cloud-edge-thing collaborative EC scenarios.This thesis con-siders the delay sensitivity of the user terminal and the constrained computing resources of the ECS and establishes a non-convex optimization model with the goal of maximizing system benefits.We transform the non-convex optimization model into a Generalized Ge-ometric Planning(GGP)problem through equation transformation.We then redesign the Branch and Bound iterative algorithm through constant lower bound relaxation and pre-pruning methods.Finally,we verify the effectiveness of the proposed algorithm through simulation experiments and improve the system's revenue.Second,given the high energy consumption overhead of current edge caching strate-gies,and the mismatch of cached content on ECS near users with their interests and prefer-ences,we propose a joint optimization strategy based on task request and content caching in the edge collaborative caching scenario.This thesis considers the delay sensitivity of the user terminal and the storage resource constraints of the ECS and establishes a non-convex optimization model with the goal of maximizing system benefits.The original problem is decoupled into the request receiving decision sub-problem and the content caching deci-sion sub-problem.We propose an iterative algorithm based on convex programming and simulated annealing to solve these two sub-problems.Finally,we verify the convergence and effectiveness of the proposed strategy through simulation experiments.Experiments point out that the proposed algorithm improves system revenue and content hit rate,and reduces the average completion time of user requirements.