The Research On Global Performance Optimized Mobile Offload Decision-making Model

With the emergence of smart phones, mobile devices, computing power hasincreased significantly, and people for the complexity of mobile applications is alsoput forward the corresponding high demand, and increase the speed of the latter ismuch higher than the former to improve the speed and solve the two do not match thedevelopment of mobile computing has become a must overcome. Mobile devices andcomputing power to solve the people’s requirements for mobile applications does notmatch the complexity of the problem, there are two solutions directions, one mobiledevice to overcome the problem of limited resources, this area is mainly to improvethe hardware to solve two more for the efficient use of existing limited resources tobetter meet the increasingly complex demands of mobile applications, mobile loadtransfer is one of the more typical approach.Offload decision-making strategy is to move the load transfer in the central partof the model, but because most of the current static load transfer decision making, theability to respond to changes in resources and the environment is weak, can notguarantee the effectiveness of real-time decision making.To solve the above problems, this paper presents a dynamic performanceoptimization based on global decision-making strategies, dynamic decision strategy ofthis paper consists of two parts:（1）For the analysis of the effectiveness of the decision-making results in theimplementation of the application process, resource monitoring module periodicallycall to get the current resource information, analyze the effectiveness of currentdecisions, wrong decisions once found, promptly corrected.（2）For the analysis of the effectiveness of the decision-making related to thecalculation model by analyzing the actual implementation of the results of theprevious analysis of the effectiveness of the decision-making related to the currentmodel (application execution time estimation models and applications to performenergy estimation model), once found the current the results of the calculation modelerror is large, the time to change the calculation model, in order to ensure theeffectiveness of subsequent decisions.（3）For a complete execution of the application path, and total execution timeanalysis of the total energy history of the execution path according to the individual needs of the optimal decision based on the analysis of objective function, under theinfluence of a decision.