Research On Computing Offload Technology In Multi-user And Multi-MEC Server Scenario

The computing offload technology in mobile edge computing can solve the deficiencies of user’s device in terms of storage resources,computing capabilities,and battery life.Execution delay and energy consumption of the mobile device to perform the task are the main factors that affect the computing offload decision.However,the transmission power of a mobile device will affect the execution delay and energy consumption for device to perform tasks.Therefore,how to optimize the network channel resource allocation and the transmission power of the mobile device and realize reasonable computing offload among multiple users under the condition of time delay constraint is one of the key challenges faced by MEC computing offloading.Based on the analysis and research of mobile edge computing technology,this thesis studies the network interference among multiple users and the matching relationship between users and MEC servers and formulates reasonable offload strategies,in order to improve the efficiency of mobile devices in performing complex computing tasks and make each user within the coverage of the MEC server have an experience with ultra-low latency and high battery continuation.Based on the in-depth study of computing offload technology in the multi-user and multiMEC server scenario,this thesis proposes a multi-user and multi-MEC server computing offload framework,focusing on the establishment and solution of computing offload model.First of all,the calculation method of energy consumption,delay and data transmission cost when tasks are executed locally and on the MEC server is given.Assign weight coefficients to the user’s task execution delay and energy consumption,take the delay and energy consumption as the user’s total cost,and minimize the total cost of all users in the system as the optimization goal.Taking the computing offload decision and the transmission power of the mobile device as optimization variables,a computing offload model in the scenario of multi-user and multiMEC server is established.Then,a computing offload and transmit power alternate optimization algorithm is designed.The algorithm decomposes the model solving problem into two sub-problems: solving the computing offload decision and solving the transmitting power.The user-MEC server matching based on the KM algorithm and the user-subchannel matching based on the delayed acceptance algorithm are used to solve the computing offload decision of each mobile device,and the low-time complexity binary search algorithm is used to optimize the transmission power.The two sub-problems are solved iteratively and alternately to find the optimal computing offload scheme.This algorithm comprehensively considers the user’s demand for delay and energy consumption and the interference among users in the sub-channels to reduce the total cost of all users in the system.Finally,a simulation experiment is designed to compare the computing offload and transmission power alternate optimization algorithm designed in this thesis with other related methods.Experimental results show that the algorithm can significantly reduce the total cost of the system and show better performance.