针对同步多链路Wi-Fi网络资源分配复杂度高并且难以同时优化系统吞吐量和传输时延以提升网络性能的问题,提出了一种基于双重深度Q网络(double deep Q network,DDQN)的Wi-Fi网络同步多链路资源分配和优化算法.将各个链路信道划分为多种不同规格的资源块(resource unit,RU)组合,利用DDQN进行RU组合选取,结合KM(Kuhn-Munkres)算法为设备分配RU,从而降低资源分配复杂度;针对网络性能提升问题,将满意度定义为吞吐量和时延的函数,联合优化资源分配和传输时长,以提升满意度,从而提升系统吞吐量、降低传输时延.在饱和业务流量及低负载流量模型下进行仿真分析,结果表明,所提算法具有较好的收敛性能,并且在提升满意度和系统吞吐量、降低传输时延方面优于对比算法.
DDQN-based resource allocation and optimization for synchronous multilink Wi-Fi networks
Aiming at the problems of high complexity of resource allocation in synchronous multilink Wi-Fi networks and the difficulty of simultaneously optimizing system throughput and transmission latency to improve network performance,a double deep Q-network(DDQN)based resource allocation and optimization for synchronous multi-link Wi-Fi networks al-gorithm is proposed.Each link channel is divided into multiple resource unit(RU)combinations with different specifica-tions,and DDQN is employed for selection of RU combinations,while the Kuhn-Munkres(KM)algorithm is adopted to al-locate RU for devices,thus reducing the resource allocation complexity.For the network performance enhancement prob-lem,satisfaction is defined as a function of throughput and latency,and resource allocation and transmission duration are jointly optimized to enhance satisfaction,thus improving system throughput and reducing transmission latency.Simulation a-nalysis is performed under saturated service traffic and low load traffic models,and the results show that the proposed algo-rithm has better convergence performance and outperforms other comparative algorithms in improving satisfaction and system throughput and transmission latency reduction.
万方数据
