Research On UAV Deployment And Task Scheduling In Mobile Edge Computing

At present,with the continuous development of Mobile Edge Computing(MEC)technology and the maturity of Unmanned Aerial Vehicle(UAV)technology,UAVs are being widely used in mobile edge computing scenarios,bringing numerous advantages and innovations to the field.However,there are still some unresolved issues in current research.Firstly,the current research lacks specific studies that consider optimizing both the number and position of UAVs.Secondly,there is a lack of research on the joint optimization of task scheduling and UAV deployment.Finally,unreasonable deployment and scheduling strategies may result in wasted UAV capabilities and increased communication latency and energy consumption,thus hindering the realization of the low latency and high bandwidth advantages of edge computing.To address these challenges,this thesis focuses on the following research contents and innovations:(1)An edge computing deployment model based on multiple UAVs is proposed to address the problem of high total energy consumption in Internet of Things(Io T)systems caused by unreasonable UAV allocation.A differential evolution algorithm with variable population size based on a mutation strategy pool initialized by K-Means++ is used to solve the model.In the algorithm,the population is initialized using K-Means++ to obtain the optimal initial position of UAVs.In addition,considering the limitations of fixed mutation strategies in traditional evolutionary algorithms,a mutation strategy pool is used to update the position of UAVs.Experimental results show that the proposed algorithm effectively reduces the energy consumption of the entire system.(2)A bi-level optimization framework called differential evolution algorithm with variable population size based on a mutation strategy pool initialized by K-Means++ and greedy algorithm framework is proposed to address the problems of high total energy consumption,low task completion success rate,and high latency caused by unreasonable allocation of multi-UAV deployment and task scheduling plans.In the upper-level optimization,the algorithm proposed in(1)is used to optimize the position and number of UAVs.In the lower-level optimization,an efficient greedy algorithm framework is proposed to optimize task scheduling.Finally,the experimental results confirm the effectiveness of the upper-level optimization,the lower-level optimization and the bilevel optimization,and reduce the system energy consumption while ensuring low delay and high task completion rate.(3)A UAV deployment and task scheduling system was built in the MEC scenario.Based on previous work,a prototype system was designed using technologies such as Java,MATLAB,and My SQL.This system can determine the number and optimal position of UAVs,and can coordinate task assignment among multiple UAVs to ensure that each UAV receives suitable tasks and completes them within the scheduled time.The system guarantees efficient task execution while also maximizing the use of UAV energy,thus improving the performance and efficiency of the entire MEC system and enhancing its reliability and robustness.Additionally,a visualization interface was created to allow the administrator to monitor the system’s smooth operation.