Research On Joint Optimization Of Computing Offloading And Resource Allocation For Multi-device And Multi-server

With the continuous and rapid development of Internet of things and wireless communication technology,various intelligent devices and novel applications are constantly emerging,which makes the amount of data in the network grow exponentially.On the one hand,the cloud computing model is difficult to adapt to the high load of network bandwidth caused by the huge amount of data transmission.On the other hand,all kinds of novel applications are characterized by high computation-intensive and delay-sensitive features.Long-distance transmission requires a large amount of energy and leads to high delay,which greatly reduces the user's service experience.To this end,edge computing has emerged as one of the effective solutions to these problems.However,the resources of single edge server are limited.With more and more devices connected to the network,operators propose to deploy multiple edge servers to provide users with continuous and reliable services.In a multi-edge server scenario,it may happen that a certain device is covered by multiple servers,and multiple devices share the resources of a certain server,which makes the edge computing environment and task offloading more complex.The inefficient task offloading strategy not only makes it impossible for users to satisfy their needs,but also leads to low resource utilization and unbalanced load at the edge.Therefore,this article focuses on the problem of computing offloading and resource allocation optimization in the scenario of multi-device and multi-server deployment.First of all,this article proposes a joint optimization strategy of computing offloading and resource allocation based on game.The weighted sum of task execution delay and energy consumption is defined as the computational overhead of task offloading,and the joint optimization problem is modeled as a problem of minimizing the computational overhead of task offloading under the delay constraint.The device broadcasts the task offloading request to the edge server within the communication range.The server that receives the task offloading request is regarded as a competitive relationship.They allocate resources according to their own resource conditions to obtain the execution of the task.Therefore,the optimization problem is changed into a non-cooperative game problem,which is called a multi-edge server resource allocation game.In this article,the existence of Nash equilibrium is proved theoretically,and an iterative algorithm is designed to find the Nash equilibrium and obtain the optimal offloading strategy of the task.Simulation results show that the algorithm effectively decreases the system computational overhead and energy consumption,improve the server resource utilization and achieve server load balancing.Secondly,this article proposes a joint optimization strategy for computing offloading and resource allocation based on multi-edge collaboration.Considering that edge servers can cooperate with each other to process tasks,there are two execution modes of tasks: local edge server and assistant edge server.Their respective execution delay model is established.Considering the fact that operators provide computing resources in a paid way,this article establishes the user service payment model.Then,the product of task execution delay and user payment service fee is defined as user overhead,and the joint optimization problem is expressed as a group of user total overhead minimization problem under delay constraint.Since the problem is a mixed integer nonlinear programming problem,the problem is decomposed into two sub-problems: computing offloading and resource allocation.The resource allocation problem is solved by convex optimization method.Then,a computing offloading and resource allocation algorithm based on improved particle swarm optimization is proposed to obtain the optimal offloading strategy of a group of tasks according to the amounts of resources allocated.Simulation results demonstrate that the algorithm has good performance in achieving server balance,effectively allocating computing resources,reducing user overhead and task execution delay.