Research On Task Offloading Strategy In Mobile Cloud Computing

With the rapid development of mobile smart terminals such as smart phones and tablet computers,relevant smart terminal services(e.g.mobile payment,mobile games,and mobile travel)have brought great convenience for people.However,its performance,such as weak CPU processing capacity,small storage space,and short battery life time,has caused many inconveniences for the smart business experience.The mobile cloud computing technology can transfer the complex tasks of the mobile intelligent terminal to the cloud processing through the network transmission,so as to make up for the abovementioned defects in the performance of the mobile intelligent terminal,and also greatly improve the task processing efficiency and users' quality of service.At present,the existed research about cloud computing offloading strategy focuses on energy consumption and delay optimization issues,but it ignores the offloading cost required while tasking offloading.To solve the problem.This thesis establish a task offloading queue model based on Lyapunov optimization method.In order to optimize the task offloading cost,a task offloading strategy that balances network delay and offloading cost is proposed.This algorithm makes full use of the dynamic effect of sending buffer queue of intelligent terminal device and makes real-time scheduling decisions according to the users' network link status and the backlog information of its buffer queue at the current scheduling time.The simulation results show that the task offloading cost is optimized while the task offloading delay is reduced.Then,considering impact of the heterogeneous network,different connections(e.g.cellular and WIFI)on the transmission broadband cost of task offloading and the effect of system offloading based on the offloading queue model.Firstly,a queue is built to model the mobile users' workload offloading and Lyapunov optimization is used to make trade-off between the offloading utility and the queue backlog.Then Lagrangian optimization method and a multi-stage stochastic programming method are respectively proposed for the deterministic WIFI connection and random WIFI connection to decide the optimal offloading workload.The simulation results show that the offloading utility can be effectively improved,and the backlog of queues for users' offloading tasks is also reduced.In addition,the utilization of cloudlets and smart terminal local resources in mobile cloud computing is not balanced.When there exist multiple user task offloading requests and multiple cloudlets platforms in the network environment,if the resource of some micro clouds is insufficient,and blindly obtaining the offloading service from the cloud center may cause longer transmission delay.In order to solve this problem,this thesis proposes an optimal matching task offloading strategy to minimize energy consumption.In order to optimize the energy consumption for offloading,this thesis proposes an optimal matching task offloading strategy based on minimum energy weight.In order to optimize the energy consumption for workload offloading,the best matching results between the user and the micro cloud are obtained.The simulation results show that user can reduce the energy consumption user of tasks offloading and response time request for the offloading task in the network through reasonable selection,so that the user can obtain an efficient offloading service.