Research On Task Assignment Scheme In Mobile Ad Hoc Cloud

The rapid development of wireless networks and mobile devices has brought great convenience to daily life.People can meet various practical needs through mobile terminals.And the mobile cloud computing technology can effectively solve the problem of limited in computing resources and processing capacity.However,there are still some problems.For example,the cloud server cannot be found,or the connection is unstable due to the distance of the remote cloud is too far.Therefore,relevant scholars have proposed the mobile Ad Hoc cloud that can be quickly networked without infrastructure.Users can complete their own computing tasks without having to access the remote cloud,by moving the shared resources of other mobile devices in the Ad Hoc cloud.In recent years,a lot of researches have been carried out by relevant scholars.However,the research on agent incentive mechanism is not deep enough.For this reason,this thesis proposes a joint optimization scheme for task allocation and load balancing based on distributed game.The scheme first uses the buyer/seller game to analyze the client and agent terminal respectively.According to the utility and cost generated during the task allocation process,the objective function of optimizing the system payoff is proposed,and then the optimal strategies of the buyer and seller are obtained through iterative update algorithm.At the same time,the strategies effectively incent the agent terminal through the game of buying and selling.For the random mobility of nodes,this paper adopts a multi-stage stochastic programming method to make modeling for communication costs,so that during the task allocation decision,we can reduce the communication costs as much as possible while optimizing system payoff.For the case that the new arrival task cannot find the suitable resource for allocation due to the excessive number of clients and the shortage of computing resources,this paper also proposes an agent terminal load balancing mechanism.Based on the game results,the computing resources in the system are dynamically adjusted through the multiple rounds sequential auctions,so that the high-load agent can release the occupied resources and accept the new task.The simulation shows that the scheme can optimize the system payoff of the mobile Ad Hoc cloud in the task allocation process,reduce the communication cost,and improve the performance of resource utilization and task success rate.Moreover,task allocation still needs to consider the system energy consumption and delay.In order to achieve better performance,this paper also proposes a joint optimization scheme of system energy consumption and task delay in task allocation.By modeling the energy consumed for computing,energy consumed for communication and task delay,it can perform pre-estimation processing according to the task assignment decision algorithm so as to select a better agent to allocate the task.And in this way,we can reduce the total system energy consumption and task delay as much as possible.Finally,the simulation verified the effectiveness of the scheme.