A Research On The Energy Consumption Of The UAV Based Mobile Edge Computing System

With the development of the Internet of Things,the performance of terminal equipment continues to improve and the number of devices accessed on the internet brings about the exploding.Moreover,the new applications based on big data and artificial intelligence,represented by video,virtual reality(VR),online healthcare,consume lots of network traffic and information resources.To better cope with the development and demand of software and hardware,Mobile Edge Computing(MEC)as an effective supplement to cloud computing has received widespread attention.MEC is deployed in a distributed manner,sinks information resources,and builds a cloud service environment on the terminal side so that applications can get a faster response.Thus,MEC can better satisfy the requirements of terminals on aspects as performance,delay,and speed.As the research further develops,Aerial-Mobile Edge Computing(A-MED)has gradually attracted widespread attention where UAV is used to build and deploy Aerial-Mobile Edge Computing(A-MED)platform.The UAV is economical,convenient,and flexible,which can better make up for the weakness of the base station.Therefore,the UAV mounted with MEC servers served as a tool for data collection,auxiliary calculation,service coverage,which is widely applied in disaster relief,public transportation,security protection,and other fields.However,it is worth noting that the UAV itself also has many limitations in terms of load,endurance,and energy consumption.In particular,the single UAV-enabled A-MEC system is more sensitive to energy consumption for running out of energy will harm the system security.Aim to tackle the problem mentioned before,this work investigates the energy consumption problem at both terminal devices and the UAV in the proposed A-MEC system.This paper proposes a hybrid offloading schema based on computing and caching operations which takes the energy consumption of various task operations into consideration.As the formulated problem is a mixed-integer nonlinear programming problem,we proposed an iterative algorithm that utilizes the successive convex approximation(SCA)technique and the block coordinate descent(BCD)method to isolate the integer variables into different subproblems.Moreover,the branch and bound(B?B)and penalty-based solutions are applied to tackle the integer programming(IP)problems in the terminal energy consumption problem and the UAV energy consumption problem respectively.The experimental results show that the branch and bound method can better find high-quality bounds in the case of low dimension,while the penalty function algorithm can quickly solve large-scale problems.