Research On Pricing Model Based Offloading Algorithm Of Mobile Edge Computing

In recent years,with the rapid development of mobile communication,network technology and artificial intelligence,a lot of new applications have been produced and the requirements for mobile terminal equipment have been further enhanced.The limited battery capacity and computing resources of mobile terminals will not be able to meet the challenges of computing-intensive and time-sensitive applications.Mobile Edge Computing(MEC)is expected to be a new paradigm to solve this problem.It deploys MEC servers at the edge of mobile networks,thus,the mobile terminal can offload the whole or part of its computing tasks to the MEC server,which reduces the time delay and energy consumption of the mobile device and improves the user experience.However,due to the limited computing resources of MEC servers,how to make the decision of offloading for mobile terminals and how to allocate and price the resources for MEC servers are key and challenging issues.In view of the above problems,the main research contents and innovations of this thesis are as follows:(1)For the scenario of single MEC server and multi-user devices of single cell,the tasks and tasks transmission of all user devices,the computing resources of MEC server,the utilities of user devices and the utility of MEC server are modeled.The concept of psychological expected price is proposed,and the reduction ratio of the actual paid price compared with the psychologically expected price is introduced into the utility functions of the user devices to measure the satisfaction degree of the user devices with the actual paid price,so that the offloading decisions of user devices will be affected by pricing.(2)Based on the above modeling,a mobile edge comptation offloading algorithm based on the pricing model in a single cell scenario is proposed.This algorithm enables user devices to make autonomous power allocation decisions,autonomously apply for computing resource requirements,and autonomously decide whether to apply for offloading and realizes the pricing mechanism of the MEC server.The offloading decisions of user devices will be comprehensively affected by factors such as wireless channel conditions,power allocation strategy,computing resource allocation strategy,and MEC server pricing.MEC server will also adjucst price according to the computing resource demands of the user devices.Specifically,First,the MEC server makes the decision using the dynamic pricing algorithm.Then,each user device makes power decisions,computing resource demand decisions and offloading decisions based on server pricing,wireless channel conditions using the Q-learning algorithm in reinforcement learning,and sends a request to the MEC server.Finally,the MEC server responds to the requests of all the user devices.The simulation results show that the algorithm can not only realize the reasonable offloading decision when the channel is fading rapidly and the price of the MEC server is changing dynamically,but also the MEC server can adjust the price according to the supply and demand relationship,and improve the utilities of user devices and the server.(3)The above algorithm is extended to the multi-cell scenario,and a mobile edge compututation offloading algorithm based on pricing model is proposed in the multi-cell scenario.Considering the wireless channel interference problem in the multi-cell scenario,an empirical interference matrix is designed in the user devices.In addition,considering the bidding problem between multiple MEC servers,a further price adjustment mechanism is designed in the MEC servers.Simulation results show that the algorithm can also improve the utilities of user devices and servers.