Research On Resource Scheduling Technology For Mobile Cloud

With the popularity of smart mobile devices and the evolution of cloud computing technology,mobile cloud computing has played an important role in our lives.It can solve the problems with mobile terminals,such as poor computing performance and small storage capacity,but faces many difficulties in resource management as the number of users grows.Research on mobile cloud resource scheduling and management technology is therefore necessary.Based on the analysis of the relevant needs and key technologies,the contributions of this work are presented as follows:(1)To solve the problems with mobile cloud data centers such as high energy consumption and low resource utilization,we improved the energy consumption model of mobile cloud data centers,and designs virtual machines(VM)with minimum energy consumption,maximum utility,and minimum number of servers as scheduling targets.An adaptive VM scheduling algorithm VMSA-IPSO for mobile cloud computing is proposed to solve the model,and the effectiveness and performance of the algorithm are verified in simulation and comparison experiments.(2)To solve the problems of low accuracy of resource prediction and high degree of service violation in mobile cloud computing,the LSTM-RNN prediction model is applied to predict the CPU resource requirements and network bandwidth changes in cloud data centers.To improve the accuracy of the prediction model,an LSTM-RNN hyperparameter optimization algorithm based on Improved Particle Swarm Optimization(IPSO)is designed.A VM scheduling algorithm VMSA-LSTM-RNN is proposed based on the prediction model.Considering the relationship between the current and future resource requirements and VMs to be migrated,this algorithm is designed to improve the migration time and overall deployment strategy during scheduling.And the effectiveness and performance of the algorithm are verified in simulation and comparison experiments.Finally,according to the requirements of mobile cloud computing resource prediction and scheduling and the application environment of the joint combat command system,a prototype system for mobile cloud computing resource prediction and scheduling was designed and initially implemented to realize the sharing of resources and information of various combat units and improve the efficiency of resource use.And the performance of the prototype system is tested and analyzed in experiments.The test results show that the function and performance of the prototype system developed meet the design requirements.