Serial Divisible Load Scheduling Algorithm For Energy Consumption-Delay Tradeoff In Chain Topology Wireless Networks

With the rapid development of research in emerging transportation fields such as unmanned driving,unmanned vehicle formations on highways constitute a chain topology wireless network determined by linear roads.There may be high-speed data interaction between vehicles.Although the data scale of this kind of computing-intensive task is very large,it has the characteristics of parallel processing and arbitrary division.The divisible load scheduling model distributes tasks to network nodes according to the processing and communication capabilities of nodes in the network.The dividable scheduling model can obtain the analytical solution of the optimal solution.Due to the positional relationship of nodes in the chain topology wireless network,the execution order of applications such as inter-node communication,task scheduling,and instruction issuance is limited in the chain topology wireless network dividable load scheduling model.Since the existing chain topology networks based on the dividable load theory do not consider the energy issue,Aiming at how to ensure the real-time performance of computing-intensive tasks in the chain topology wireless network and improve the system energy utilization,this paper studies the energy consumption-delay tradeoff optimization problem of the chain topology wireless network serial dividable load scheduling based on dynamic programming.In this paper,the following work is carried out for the two cases of the dividable load model with return and without return:(1)This paper proposes a serial dividable load scheduling algorithm with chain topology wireless network energy-delay balance.Assuming that the chain topology wireless network node supports dynamic voltage and frequency scaling technology,the node selects the computing and communication capability level based on the dynamic programming theory according to the maximum computing delay that can be tolerated by the computing-intensive task,and then divides the task,according to the communication mode of chain topology,the sub tasks are distributed to other nodes in the network in turn.Finally,based on the theory of divisible load,the optimal load distribution strategy under the current computing and communication capacity is obtained.(2)In order to calculate the allocation strategy conveniently in large-scale network,Markov chain model is used to analyze the energy consumption-delay tradeoff serial divisible load scheduling process and its return situation in chain topology wireless network,which is equivalent to queuing model.Then Little theorem is used to analyze the total task completion time in the process of serial divisible load scheduling.The simulation results show that the algorithm can adjust the voltage and frequency of wireless network nodes to the most suitable level under the premise of meeting the total task completion time constraint,so as to reduce the total energy consumption in the load scheduling process.At the same time,the equivalence of the chain topology serial divisible load model and Markov chain model is verified.