【作者】 陈小惠；

【导师】 程文青；

【作者基本信息】 华中科技大学 ， 通信与信息系统， 2013， 博士

【摘要】 随着无线应用的不断涌现和业务需求的与日俱增,无线频谱资源的稀缺性与当前无线通信系统频谱利用率低下的矛盾日益凸显,严重制约着无线通信技术的进一步发展。针对频谱资源紧缺现状,无线研究领域提出了多种解决方案,分别从时域、空域和频域等多个方面提升频谱资源的利用率。在众多的解决方案中,可变频宽技术易于实现,能够推广应用于现有通信系统。因此,可变频宽无线网络的研究具有重大的学术意义以及广泛的应用前景。可变频宽技术作为一种新的资源调控手段,能够更加灵活地改变网络运行状态和性能,但同时也增加了信道划分的维度,导致无线网络资源分配更加复杂。本文以提升IEEE802.11网络性能为目的,结合可变频宽技术对无线网络中的资源分配问题进行研究。论文首先对可变频宽技术的实现方法进行了概述,并分析和总结了IEEE802.11网络资源分配的研究问题和现状,然后对其中三个典型的网络场景中的资源分配问题分别展开探讨,最后分别设计了相应的资源分配算法,并通过大量的仿真实验验证了算法的性能。本文的主要工作包括以下几个方面：(1)在合作的单跳接入网络场景中,信道分配和用户关联是资源分配的两个主要问题。考虑信道频宽对无线链路通信距离的影响,论文首先分析了信道分配与用户关联的相互影响,指出信道分配与用户关联应当联合优化处理；接着,以最大化系统吞吐量与网络公平性两个系统性能为优化目标,建立了联合优化问题的数学模型；最后,基于粒子群优化框架提出了信道分配和用户关联算法。仿真实验结果验证了联合优化机制能够显著提升网络性能。(2)在非合作的单跳接入网络场景中,由于用户的自私性,用户以最大化个体接入收益为目的选择不同的无线接入点。考虑信道频宽对接入性能的影响以及用户QoS要求,论文首先设计了可变频宽无线网络中的用户效用函数；然后在用户有限理性的假定下,基于演化博弈理论建模了无线接入点选择问题；最后设计了一种基于复制者动态的分布式学习算法,引导用户得到此问题中的演化稳定策略。仿真的实验验证了本文所提出的分布式算法能够有效地收敛到演化均衡。(3)在多跳无线网状网络中,所有路由节点统一部署且具备中心调度的能力,信道分配和路由是此场景下资源分配的两个主要问题。论文首先建立了可变频宽无线网络中的网络连通性模型和累积干扰模型；然后考虑信道分配与路由的耦合关系,建立了联合优化问题的数学模型；最后基于列生成优化框架对联合优化问题进行解耦,并设计了一种基于对偶价格的贪心式算法求解信道分配策略。仿真实验结果表明本文所提出的算法能够有效求解信道分配与路由联合优化问题。本文的研究工作对可变频宽无线网络中的资源分配问题进行了较为全面的分析和研究,并分别设计了相应的资源分配算法和解决方案,能够为可变频宽技术的应用及无线网络资源分配优化设计提供参考。更多还原

【Abstract】 Within the booming of new wireless applications and various service demands, the conflict between the scarcity and under-utilization of wireless spectrum resource in nowadays becomes increasingly critical and severely restricts the further development of wireless communication technology. To address the spectrum under-utilization problem, the research community has proposed several solutions to improve the spectrum utilization in terms of time, space, and frequency domains. Among all the solutions, variable channel-width is considered as a promising technique as it is flexible to deploy and popularize in existing communication systems. Therefore, in this work, we mainly focus on this technique and investigate some of the key issues and its practicability in future applications.As a new resource allocation mechanism, variable channel-width technology provides a more flexible way to adapt the network status and system performance. However, it also complicates the network resource allocation as the variable choice of channel-width opens another dimensionality to the resource allocation problem. In the IEEE802.11wireless network, our work aims to improve the network performance by employing the variable channel-width technique in our resource allocation problem. In this thesis, we first give a brief introduction to the variable channel-width technology, and then summarize some related works about resource allocation in IEEE802.11wireless network. Then, we focus on the resource allocation problems in three typical network scenarios, and design corresponding resource allocation algorithms. In the final, we validate these algorithms through extensive simulations.Main contributions of this thesis consist of the following aspects:(1) In the scenario of cooperative single-hop access networks, the channel allocation and user association are the two main problems in resource allocation. Considering the influence of variable channel-width on the communication distances of wireless links, we first analyze the interaction between channel allocation and user association and note that a joint optimization is required between these two problems. Then, we formulate a joint optimization model with the objective of maximizing system throughput and network fairness. Moreover, to solve this problem, we propose a channel allocation and user association algorithm based on the particle swarm optimization framework. Simulation results show that the joint optimization model can improve the network performance significantly.(2) In the scenario of non-cooperative single-hop access networks, selfishness urges different users to choose the nearby wireless access points to maximize their individual utilities. We first design a new user utility that incorporates the variable channel-width, considering its impact on the access performance and user’s QoS requirement. Then, under the assumption of limited rationality, we formulate the access point selection problem in an evolutionary game framework. Moreover, we propose a learning algorithm based on the replicator dynamics to lead users to achieve an evolutionary stable strategy in a distributed way. Simulation results show that the proposed algorithm can converge to the evolutionary equilibrium efficiently.(3) In the scenario of multi-hop wireless mesh networks, all wireless routers are deployed by the same authority capable of centralized scheduling. Channel allocation and routing become the two main problems in such scenario. In this thesis, we first improves the network connectivity model and aggregate interference model to account for the variable channel-width in wireless network, and then propose a joint optimization framework to address the resource allocation problem considering the coupling between channel allocation and routing. We employ the column generation method to decouple the joint optimization problem, and propose a greedy algorithm to search for the channel allocation strategy based on dual price. Simulation results show that the proposed algorithm can solve the joint optimization problem efficiently.This work presents a comprehensive study on the resource allocation problem with variable channel-width in the wireless network, and design several resource allocation algorithms corresponding to different network scenarios. Our analysis and simulation results provide a substantial reference for the application of variable channel-width technology and the design of resource allocation algorithms for the future wireless technologies.更多还原