Research On Task Offloading Strategy In Ad Hoc Cloud

The application of mobile cloud computing technology provides an effective way for mobile devices to break through their own resource limitations.However,in many scenarios,there are no available default infrastructure to ensure that users can access anytime,or large transmission delay due to far away servers.In the ad hoc cloud environment,the mobile device which has limited resource can offload computationintensive tasks on other mobile devices with relatively rich resources to implement the expansion of its own resources.But in the actual scenario,task offloading needs to consider the random mobility and heterogeneity of the nodes.Based on the above analysis,aiming at the task offloading problem in the ad hoc cloud environment,this thesis first analyzes and studies the characteristics of ad hoc cloud and the multi-influence factors in the process of task offloading,and on this basis constructs a system model that conforms to the actual scenario.Then,in order to improve the offloading efficiency,a multi-criteria task allocation algorithm based on mobility prediction is proposed.Firstly,according to the time-series analysis predicts the escape time of the node and uses it as a measure of node mobility.And then the analytic hierarchy process is used to obtain the weight of CPU speed,core numbers,workload and others factor.Finally,the task assignment is based on the combined weight.The simulation results show that compared with the random task assignment algorithm and the Min-Min algorithm,the proposed algorithm can effectively reduce the task execution time and energy consumption.In addition,in order to meet the multi-objective needs of users,a cluster head node is introduced as a centralized controller to design a multi-objective offloading decision model.Tasks are prioritized according to user actual requirements,the task completion time,energy consumption and overhead are considered for offloading decision.A task offloading algorithm based on genetic algorithm and ant colony algorithm is proposed.Firstly,the feasible solution is obtained by using the global random fast search capability of genetic algorithm.Then it is used as the initial pheromone of ant colony algorithm.Finally,the positive feedback mechanism of ant colony algorithm is used to realize the optimal solution of task assignment.The simulation results show that compared with the random task assignment algorithm,the heterogeneity-aware task allocation algorithm and the genetic algorithm,the proposed algorithm can effectively reduce the task completion time and energy consumption,and the convergence speed is better than the genetic algorithm.