Research And Implementation Of Mobile Computation Offloading In Multi-user Environment

At present,mobile devices have become an indispensable part of our daily life.And developers are developing diversified mobile applications with more complex functions.Due to the limitation of volume and weight,mobile devices can not meet the requirements in terms of processing capacity,storage capacity and battery capacity.Mobile Edge Computing(MEC)and computation offloading technology use the server's efficient computing power and resources to complete tasks by offloading complex computing tasks to nearby servers.By computation offloading,you can extend battery life and improve mobile device performance.The existing computation offloading decision-making scheme achieves the purpose of reducing mobile device energy and application delay by parting mobile programs under different granularities such as applications,modules and functions.These algorithms are very complex and require a lot of solution time for large-scale applications.Therefore,this thesis proposes a computation offloading algorithm based on maximum flow minimum cut,selects a reasonable program partition granularity and designs a program decomposition algorithm to complete the computation offloading.In a multi-user scenario,due to the limited resources of the MEC server and the heterogeneity of user requirements,the MEC server cannot meet the needs of all users.This thesis proposes a resource scheduling algorithm based on auction theory in a multi-user environment to meet user needs and maximize the benefits.The main work and contributions of this thesis are as follows:This thesis proposes a computation offloading algorithm based on maximum flow minimum cut.Firstly,the objective function of the mobile program computation migration is given by defining the network diagram model of the mobile application.The program partition problem is transformed into the minimum traffic minimum solution problem of the network diagram.Then,the algorithm is accelerated by simplifying and compressing the network.The network is divided into sub-networks based on the label propagation algorithm,the compression conditions are defined,and the network graph is compressed.Finally,the maximum flow of the new network graph is solved by the pre-flow advancement algorithm with the minimum cut of the maximum flow,and the minimum cut is solved.According to the calculation result,the optimal solution of the mobile application's computation offloading is obtained.Through simulation experiments,it is proved that the algorithm can optimize the delay and energy consumption,and speed up the algorithm,and it can also have good effects on large-scale networks.This thesis proposes a resource scheduling algorithm based on auction theory in a multi-user environment.This thesis establishes an auction model for the auction scene of multi-user and multi-MEC servers.Under the constraints of computing resources and wireless resources,the objective function of the matching problem between mobile users and MEC servers is given.The MEC server priority vector of the mobile user pair is designed according to the dynamic resource indicator of the MEC server.Users bid based on priority.And use the ant colony algorithm to solve the winner and match the user to the appropriate computing migration MEC server.In order to enable the service network resources to be allocated according to the user's real needs,to avoid malicious bidding or virtual bidding,and to ensure the optimal profit of all users,the auction payment rules are designed based on VCG theory.Through simulation experiments,it is proved that the algorithm in this thesis can respond to more users and bring higher profits to the system.