【作者】 吴凡；

【导师】 毛玉明；

【作者基本信息】 电子科技大学 ， 通信与信息系统， 2015， 博士

【摘要】 长期以来,无线通信网络设计主要聚焦在如何在有限的通信频带上提供更快的数据传输速率以满足用户日益增长的带宽需求。伴随无线通信网络的迅猛发展,巨大的能耗开销已成为不容忽视的问题。因此,以降低能耗,减少碳排放量为设计目标的绿色网络已成为新一代无线通信网络的发展趋势。而作为无线通信网络核心技术之一的无线资源分配技术势必也将在节能减排方面起到关键性作用。为了提升频谱效率,诸如正交频分多址接入,载波聚合,中继等大量先进的通信技术被无线通信网络所应用。而大量新技术的应用使得新一代无线网络资源呈现出维度更高,粒度更细的特性。这既为节能设计提供了可能,同时也为无线资源分配引入了更多新的、复杂的约束条件,提出了新的挑战。不仅如此,能量供给方式的变化,特别采集能量的使用,是无线网络资源分配需要面对的另一个新课题。随着能量采集技术的发展,网络的能源既可以是采用传统的电网供电,也可以使用环境能量(风能,太阳能,振动能等)。而采集能量的使用将使得未来网络可以实现自给供能,持久运行,但与此同时,无线资源分配技术也必须解决由采集能源而带来的能量供给的随机性问题。综上,本学位论文以绿色节能作为无线网络资源分配的优化目标,以多维度无线资源在使用上的相互制约关系为优化约束条件,采用跨层优化的设计方法,以非线性规划为理论基础,研究绿色无线网络资源分配中的若干关键问题。拥有载波聚合能力的OFDMA无线网络具有典型的多维度,细粒度特性,可供分配的资源不仅包括二维时频资源块(RB),发射功率,还包括分量载波(CC),多维度资源给分配引入了复杂的约束关系,这不仅体现在同类型资源的分配上,而且不同类资源的使用也会存在相互制约。本论文首先研究具有资源高度耦合特点的OFDMA无线网络的建模、分析和求解的方法,作为后续节能设计的基础。针对多种载波聚合能力用户混合接入的网络场景,本论文以最大化网络效用(NUM)为目标,建立起CC,RB以及功率的联合分配模型。区别于现有研究,本论文提出了一种动态频谱资源指派与自适应功率分配相结合的算法(JCRPA算法),该算法以迭代的方式运行,充分挖掘了系统无线分集增益。而为了减小CC切换带来的系统开销,在JCRPA算法的基础上,进一步设计了CC切换控制机制。以能效为优化目标的资源分配技术是本论文研究的重点。由于能效目标函数不再是凹的,这使得在谱效优化中广泛采用的凸规划理论无法直接应用于对能效模型的分析与求解之中。但是,能效目标函数所具有的广义凹性却能够确保优化问题的KKT条件是最优解的充要条件,这为理论分析能效资源分配问题提供了基础。然而,现有研究大多着眼于利用KKT条件的必要性去分析和求解问题,而缺乏对其充分性的利用。本论文则充分利用KKT条件的充分性,提出了能效注水法,该方法通过构造满足KKT条件的决策变量和拉格朗日系数进而获得闭合最优解,为求解最大化能效的OFDMA资源分配问题提供一条新的有效途径。利用能效注水法,提出了子载波与功率联合分配算法(JERA算法),该算法以较小的计算复杂度实现了通过子载波指派与功率分配的联合优化,并在满足用户最低速率要求的前提下,实现了系统效用的最大化。本论文将OFDMA网络无线资源分配模型进一步扩展到支持中继转发的通信系统之中,研究了以能效为优化目标,同时满足基站与中继节点各自最大发射功率约束的资源分配问题。该分配模型属于分式规划,而由于能效函数的广义凹性,使得这类分式规划可以高效求解。基于凸优化与非线性分式规划理论,并结合译码转发中继通信的特性,本论文设计了一种能效优先的绿色资源分配(GRA)算法,实现了对子载波通信模式选择与用户调度,以及基站与中继节点功率控制。能量采集将深刻影响未来无线通信网络的设计。而从资源分配的角度分析,首先需要解决的难题是如何使功率分配满足因使用采集能量而给系统引入的新约束:能量因果约束和电池容量约束。这些约束使得在许多分配模型的求解中行之有效的“注水”功率分配方法不再适用。而由于该问题的KKT条件也具有充分性,所以沿用在能效注水法中所采用的构造最优解的思路,本论文提出了一种采集能量分配的新方法——能量转移法。该方法通过采用转移能量,首次得到了满足因果约束的功率分配表达式,并给出了最优转移能量的充分条件。根据该充分条件,本论文得到了最优功率分配所满足的转移能量方程组,并给出了该方程组解的解析表达式。将转移能量最优条件进一步与方程组扩展与收缩变换相结合,本论文提出了构造最优转移能量线性方程组的方法。并在此基础上,设计了能量转移功率分配(ETPA)算法以较低计算复杂度获得了最优功率分配的准确值。更多还原

【Abstract】 During the past decades, the design of wireless communication networks has been focused on how to improve the speed of the data transmission over the limited bandwidth to meet the growing bandwidth requirement of users. With the explosive growth of wireless communication networks, the tremendous energy consumption is becoming a serious problem which cannot be ignored. Therefore, the green network with lower energy consumption and carbon footprints is a trend for the evolution of the future wireless communication networks. As one of the key technologies in wireless communication networks, radio resource allocation certainly will play an important role in energy saving and emission reduction.In order to improve the spectral efficiency, many advanced technologies such as orthogonal frequency division multiple access, carrier aggregation, and relaying have been applied in wireless networks. The application of the new technologies makes the radio resource have higher dimension and finer granularity. Not only does it open the door to energy efficient design, but it also imposes more new and complicated constraints and hence making the problem more challenging. Moreover, the change of power supply, especially the utilization of harvested energy, is another issue to be faced in the radio resource allocation. With the development of energy harvesting technology, networks can be powered by not only the power grid but also the ambient energy, such as wind energy, solar energy, vibrational energy, and etc. The future wireless networks powered by harvested energy are energy self-sufficiency and perpetual operation. However, the randomness of the available harvested energy is a new problem to be solved. Hence, in this dissertation, applying the cross-layer design method, under the guidance of nonlinear optimization theory, the radio resource allocation problem with the objective of energy saving, subject to the multidimensional constraints is intensively studied.The OFDMA wireless networks with carrier aggregation are the typical networks wherein the resources have multiple dimensions and fine granularity. The radio resources including two-dimensional resource blocks, transmission power, and component carriers impose much more complicated constraints on the allocation of both homogeneous resources and heterogeneous resources. In this dissertation, the methods of modeling, analyzing, and solving for the tightly coupled resource allocation problem in OFDMA networks is first investigated, which are the important foundations for the energy-efficient design. The joint CC, RB and power allocation problem is formulated as a mathematical programming with the objective of the Network Utility Maximization(NUM). Distinguished from many existing methods, our approach, i.e., JCRPA algorithm, combines dynamic spectrum resources assignment and adaptive power allocation technology. With the iterative update, the wireless diversity of the system can be fully exploited. Furthermore, with the cooperation of the JCRPA scheme, a CC switch control mechanism is devised to reduce the overhead due to the frequent CC switching.The focus of our research is on the energy-efficient resource allocation. Because the objective function of the energy efficiency is not concave, the convex methods which are widely used in the spectral efficiency optimization cannot be directly applied to solve the energy efficiency optimization. However, the general concavity of the energy efficiency function guarantees that the KKT conditions are necessary and sufficient conditions for the global optimal solutions, which is the theoretical basis of energy-efficient resource allocation. Different from the existing approaches utilizing the KKT conditions only as necessary conditions, we exploit the KKT sufficient optimality theorem to solve the problem and obtain the optimal solution in closed-form. Specifically, the energy-efficient water-filling approach is proposed to obtain the optimal solution by constructing the decision variables and Lagrange multipliers to meet the KKT conditions. This approach provides a new effective way to find the energy efficient solutions. Based on the theoretical analysis, a joint energy-efficient resource allocation(JERA) algorithm is proposed to maximize the energy efficiency under the under the minimum data rate requirements with significantly low computational complexity.The OFDMA-based radio resource allocation model is further extended to a relay-aided wireless communication system. The energy efficiency maximization problem is formulated as a fractional programming with the consideration of individual power constraints at base station and relay stations. Due to the general concavity of the energy efficiency function, this type of the fractional programming can be effectively solved. Based on convex optimization and non-linear fractional programming theory, and by exploiting the properties of decode-and-forward(DF) relaying communication, a green resource allocation algorithm is proposed to solve the joint transmission mode selection, subcarrier allocation and power control problem.Energy harvesting will have a profound impact on the design for the future wireless communication networks. From the views of resource allocation, the first issue to be handled is how to meet the new constraints imposed by harvested energy, i.e., the energy causality constraint and the battery capacity constraint. The additional constraints make the problem cannot be solved by using the water-filling method whose effectiveness is demonstrated in the spectral efficiency problem. Following the approach which utilizes the optimality sufficiency of the KKT conditions, a novel energy-transferring approach is proposed. By using the transfer energy, the power allocation expression which meets the energy causality constraint is first obtained, and the sufficient condition for the optimal transfer energy is also derived. Based on the sufficient condition, the transfer energy equations are provided and the closed-form solutions are obtained as well. Combining expansion and contraction of the transfer energy equations, an energy-transferring power allocation(ETPA) algorithm is proposed to obtain the accurate optimal power allocation with low computation complexity.更多还原