异构车联网基于队列稳定的最小时延资源分配算法

Queue Stable-Minimum Network Delay Algorithm for Resource Allocation in Heterogenous Vehicular Network

基于长期演进(Long Term Evolution,LTE)网络与车载自组织网络(Vehicle Ad-Hoc Network,VANET)结合的异构车联网场景中,使用队列模型对簇头消息数据包的转发问题建模,提出队列稳定前提下的最小化网络时延的资源分配算法(Queue Stable-Minimum Network Delay Algorithm,QSMNDA)。QS-MNDA通过马尔科夫不等式保证队列的溢出概率小于阈值,建立最小化网络时延优化问题,利用李雅普诺夫优化方法,动态控制多时隙下的队列稳定,并采用拉格朗日对偶分解,迭代计算出各时隙下簇头车辆对于LTE各资源块上分配的功率。仿真结果表明,QS-MNDA降低整体网络时延约10%,同时实现平均簇头队列长度下降8%左右,优化整体网络的性能。

Based on heterogenous network which combined long term evolution(LTE) and vehicle ad-hoc network(VANET),using queue model to modeling cluster head forwarding data packets,a queue stable-minimum network delay algorithm(QS-MNDA) was proposed.QS-MNDA ensure the queue stable by using Markov queuing model,and establish a minimum network delay optimization problem,dynamically control queue stability under multiple time slots by means of Lyapunov optimization method.With set up optimization problem,adopting Lagrange dual decomposition,iteratively calculates the power allocation strategy of cluster head vehicles for each resource block in LTE in each time slot.As simulation result shows,this algorithm reduces the overall network delay of about 10%,at the same time,the average queue length of cluster heads is reduced by about 8% to optimize the overall network performance.