Research On DTN-based Offloading Algorithms In Vehicular Cloud Computing Environments

In recent years,people began the study of cloud computing and vehicle network,that is vehicle cloud computing.With the sharp increase in vehicle users,real-time transmission of data between users,as well as real-time roads information sharing between users are urgent need for people.Although cloud computing can provide huge computing resources and security for the vehicle network,there are still some problems are not resolved.Tasks offloading means offloading the tasks to other terminal nodes to calculate in the case of limited resources and network,then the agent node sent the results to the destination node.Due to the characteristics of the mobile vehicle,such as the direction of mobility,high mobility,and that can be used to interrupt the communication link,making the vehicle network can't be as direct as cloud computing to offload tasks.Therefore,the effective tasks offloading algorithm under this network needs to be solved urgently.This paper mainly studies and analyzes the unloading algorithm of vehicle cloud computing.Because the vehicle network is a typical DTN network,so this paper uses the DTN algorithms.In this paper,we propose lots of offloading algorithms,which include the direct transfer algorithm of the general network,the direct transfer algorithm based on DTN and the flooding algorithm based on DTN,and the performance comparison of the three algorithms are analyzed in the experiment.After obtaining the optimal task offloading algorithm,this paper also studies the number of agent nodes with variable parameters.The number of agent nodes is related to whether the agent resource allocation is reasonable.All experiments are based on MATLAB platform,the experimental results show that the flooding algorithm based on DTN is better than the other two algorithms,the number of agent nodes is too large or too small will cause the waste of resources,and even affect the relevant performances of the algorithm.