【作者】 张昕；

【导师】 叶梧；

【作者基本信息】 华南理工大学 ， 通信与信息系统， 2010， 博士

【摘要】 随着人们对高速数据通信需求的迅速增长,无线通信系统中的频谱资源变得越来越稀缺,促使人们去开发能够有效地利用无线资源的新技术。OFDMA技术可以改善系统的频谱效率,并且能够有效对抗由于多径效应等引起的窄带干扰,因此成为宽带无线通信系统的关键技术。本学位论文的研究工作主要围绕OFDMA系统的资源分配问题展开,利用凸优化以及智能计算等技术手段,分别提出了在蜂窝网络、中继蜂窝网络以及认知网络等三种网络架构下的OFDMA系统的资源分配算法,具体的研究内容与成果如下:(1)研究了多用户OFDMA系统中基于速率最大化的RA(速率自适应)问题。提出了一种基于对偶分解的分布式资源分配方法,把一个复杂的优化问题分解为若干个关于用户的子问题进行并行求解。各个用户只需要利用自身的CSI的值以及基站提供的相关参数,就可以通过对子问题的求解获得各自的子载波以及功率的分配方案。所提算法具有较低计算的复杂度,并且能减少基站的计算负荷。仿真结果表明,所提算法能够在较少的迭代步数内得到一个近似最优解,并且与最优解的性能非常接近。(2)研究了多用户OFDMA系统中基于功率最小化的MA(余量自适应)问题。提出了一种基于蚁群算法的资源分配方案,首先将MA问题建模成一个二部图,而对原问题的求解等效为在图中寻找一条最优路径的问题,然后利用蚁群算法使蚂蚁在信息素及启发式信息的引导下构建最优的路径。由于蚁群算法具有信息正反馈以及启发式搜索等特点,使得算法可以通过较少的迭代收敛到一个近似最优解。仿真结果表明,所提算法的性能与最优值相比仅相差1dB,而与静态子载波分配算法相比性能可以提高1~2dB。(3)研究了多跳中继的OFDMA系统的资源分配。建立了一个在每个接入节点功率受限,以及满足每个用户最小速率要求的条件下,最大化系统容量的数学模型。提出一种基于对偶分解的资源分配算法,将该问题分解成若干个关于各个子载波的子问题。通过对各子问题的求解,可以获得最优的中继节点的选择方案以及功率、子载波的分配方案。仿真结果表明,该算法收敛速度较快,并且能够在保障不同用户速率需求的前提下,有效地提高系统的容量。(4)研究了解码转发中继OFDMA系统的资源分配。提出了一种基于比例公平的资源分配算法,有效地解决了系统容量与用户公平性之间的矛盾,并且采用子载波配对的方式进一步提高系统的容量。根据比例公平的准则,建立了一个在系统总功率受限的条件下,最大化各用户的速率对数和的数学模型。为降低求解的复杂度,将优化问题分为两个步骤,首先在等功率分配的情况下求出最优的子载波分配以及中继节点选择的方案,然后再根据注水定理进行功率分配。仿真结果表明,所提算法能够在保证用户公平性的条件下,有效的提高系统的容量。(5)研究了认知网络OFDMA系统中的资源分配。在实际的系统中,认知网络往往不能获得授权网络中主用户的信道状态信息的准确值,因此文中研究在非完美信道状态信息下的资源分配算法。首先将该问题建模为一个在主用户的中断概率小于限定值的条件下,最大化认知用户系统容量的数学模型。然后采用基于对偶分解的资源分配算法求出了最优的功率以及子载波的分配方案。最后通过仿真分析了非完美的CSI值对系统性能的影响,并且验证了所提算法的有效性。更多还原

【Abstract】 With the fast growing demand for high data rate communication, the spectrum resource in wireless communication systems becomes more and more scarce. This stimulates the development of new technology to achieve efficient utilization of radio resources. OFDMA is capable of enhancing the spectrum efficiency of the system. Furthermore, it can combat the inter-symbol interference causing by the multi-path effect. Therefore, OFMDA has become one of the key technologies in broadband wireless systems. This thesis focuses on the problem of resource allocation in the OFDMA systems. By utilizing the tools of convex optimization and intelligent computing, the resource allocation algorithms in three kinds of networks including the cellular network, the relay cellular network and the cognitive network based on OFDMA are proposed. The main work and contributions of this thesis are as follows:1) The problem of Rate Adaptive (RA) in multiuser OFDMA system is studied. A distributed resource allocation algorithm based on dual decomposition is proposed. The complicated problem is decoupled into several subproblems with regard to the users that can be solved parallelly. Each user can obtain its own subcarriers and power allocation scheme by solving the subproblem according to its local channel state information and the corresponding parameters provided by the base station. The proposed algorithm has a low computational complexity, and is able to reduce the calculation payload of the base station. Simulation results show that the proposed algorithm can converge quickly to an approximated optimal solution in a small number of iterations.2) The problem of Margin Adaptive (MA) in multiuser OFDMA system is studied. An ant colony algorithm based adaptive resource allocation is proposed. The problem is modelled as finding a minimum cost path in a graph. The ants’solution construction is guided by pheromone trail and heuristic information. Due to the unique heuristic searching mechanism of the ant colony algorithm, the proposed algorithm is guaranteed to converge quickly to an approximately optimal solution. Simulation results show that the proposed algorithm only loses 1 dB in comparison with the optimal solution, and gains 1-2dB in comparison with the fixed allocation scheme.3) Resource allocation in multihop relay OFDMA systems is studied. The problem is formulated as a sum rate maximization problem subject to each access node’s transmission power and each user’s minimum rate constraints. A dual decomposition based resource allocation algorithm is proposed. The problem is decomposed into several subproblems with regard to the subcarriers. By solving the subproblems, the optimal relay selection scheme, power allocation and subcarrier assignment is obtained. Simulation results show that the proposed algorithm can efficiently improve the capacity of the system, and guarantee each user’s QoS requirement at the same time.4) Resource allocation in Decode and Forward (DF) relay OFDMA systems is studied. A proportional fairness based resource allocation algorithm is proposed. The proposed algorithm performs a good tradeoff between system capacity and fairness. The system capacity is further improved by subcarrier matching. According to the proportional fairness criterion, the problem is formulated as a sum logarithmic rate maximization problem subject to total transmission power. In order to reduce the computational complexity, the original optimization problem is decomposed into two steps. In the first step, the subcarrier assignment and relay selection scheme are determined based on equal power allocation. In the second step, the power allocation is achieved by waterfilling theorem. Simulation results show that the proposed algorithm can efficiently improve the capacity of the system, and guarantee the fairness among users simultaneously.5) Resource allocation in cognitive OFDMA system is studied. In practical systems, it may not be possible for the cognitive transmitter to know the perfect channel state information value of the primary user in the authorized network. So the resource allocation with imperfect channel state information is studied in this thesis. The problem is formulated as a weighted sum rate maximization problem subjected to each primary user’s outage possibility constrains. A dual decomposition based resource allocation algorithm is proposed to obtain the optimal power and subcarrier allocation scheme. The influence of imperfect channel state information to the system is analyzed through simulation, and the efficiency of the proposed algorithm is also proved.更多还原