Uninstall Calculated Motion Prediction And Research Of Mobile Switching

With the development of the mobile terminal as smart phone and table PC, the mobile terminal has more and more functions.But CPU, memory, bandwidth,battery capacity of the mobile terminal limits the ability.With the development of the cloud computing technology, virtualization technology and WI-FI technology,it can offload the tasks of the smart phone to server with the WI-FI and3G network.The computation offloading is a technology that can reduce the energy and develop performance in mobile cloud computing,when some tasks offloaded to server.There are two main drawback in the existing computing unloading.(1) It could not consider the computing ability of the source surrogate.it will result some source surrogate is assigned too many tasks.This will produce large amounts of handover overhead.(2) It could not consider the influence of handover when the tasks are offloaded.It results in that it has not appropriate target surrogate when the source surrogate will be handover,resulting in more handover cost.For the first question,using the signal attenuation model and beacon forecast,this paper can predict the computing ability of source surrogate.It can reduce the handover cost by reduce the case of source surrogate be assigned too many tasks. For the second question,the paper proposed the offloading algorithm based predict and handover.It will the task computation-intensive and communication intensive assign the source surrogate that communication ability, communication the link resource surrogate.It will have the fit target surrogate when it will handover,that reduced the cost and developed the performance of the program.This paper proposed the predict method and the offloading algorithm based predict and handover. Though using the matlab,At first,it verified that it could develop twelve percents with the predict.Then when the number of surrogates and tasks change,the offloading algorithm based predict and handover can reduce the cost of the eight percents than the other predict algorithms.