新一代通信网络中的干扰协调与时延限制

发布时间：2018-04-21 00:19

本文选题：干扰管理 + 博弈论　；参考：《电子科技大学》2016年博士论文

【摘要】：随着基于IP(Internet Protocol)的下一代无线通信业务的迅速发展,无线网络的容量及基站之间连接的骨干网络业务调度和路由面临新的挑战。随着学术界和产业界的严密关注,在4G(4th generation)无线网络中对无线物理层进行了一系列的资源整合以提高资源利用率,如OFDMA(orthogonal frequency division multiple access),基站协作多点技术(Co MP,coordinated multi-point),CA(carrier aggregation)等。通过资源整合可以大幅度提升网络容量,增强小区边缘覆盖,但与此同时也给干扰管理带来了一系列新的重要挑战。因此,下一代无线网络里干扰协调技术呈现出新的内容,如基站多点协作中的基站协作簇形成及用户调度,基于OFDMA的基站发送功率分配等。另一方面,急剧增加的移动IP业务,尤其是5G(5th generation)无线网络物理层提出采用mmwave,massive-MIMO(massive multiple-input multiple-output),超密集小区等方案进一步支持比4G高1000倍以上的无线数据速率后,从提升资源利用率和系统能效的角度出发,如何在各种“大规模”资源消耗之间进行折衷、并在基站之间实现具有时延受限功能的业务调度,路由及无线发送是下一代无线网络所面临的另一个挑战。针对上述问题,本文结合下一代无线网络中的基站多点协作等资源整合技术,首先研究其干扰情况并提出切实可行的干扰协调方案。接着,根据无线网络物理层的实际传输能力,从系统能效的角度进一步提出了IP业务包在无线网络中的业务调度及路由算法,所提算法具有能够绑定数据流最大传输时延(worst-case delay)的特点。从无线网络干扰管理的角度,我们提出了一套分布式基站多点协作簇的自适应动态形成机制,对无线通信环境的动态变化具有很强的自适应能力。同时,分布式的决策机制简单且具有可扩展性。得到基站簇和用户调度结果后,我们在小区内所有正交频带上执行经典的WF(water-filling)功率分配算法得到基站在各频带上的功率分配策略空间。接着,我们采用纳什非合作功率博弈方法对基站协作簇中的功率分配进行建模。通过纳什博弈理论,我们证明了所提博弈模型具有唯一纳什均衡点(NE,Nash Equilibrium),并给出了寻找唯一NE点的搜索算法。性能仿真表明,所提出的基站协作方案能大大提高系统容量,且其分布式的设计既简单又具有可扩展性,因此具有很强的工程可实现性。尽管启发式算法具有很高的工程价值,但从理论研究角度,实现系统容量最大化的传输调度及功率分配联合最优解对评估启发式算法性能具有重要意义。然而该优化问题是一个混合型优化问题,现有的文献中几乎都是采用用户调度-功率分配的两步式启发式求解。本文先提出了一种新颖的PFM机制(power-fractionizing mechanism)将离散变量的传输调度与连续变量的功率分配优化问题统一到SP(signomial programming)优化模型中。基于几何规划优化方法(GP,geometric programming)给出了这类联合优化问题的局部最优解,并且从GP算法的经验上,该局部最优解往往就是全局最优解。性能仿真结果也验证了最优化算法比启发式算法具有性能上的优越性与性能表现的稳定性。在物理层资源整合的基础上,如何在下一代无线网络、尤其是5G中各种“大规模”资源消耗之间,以及有线路由与无线传输的时延之间进行高能效的折衷是新一代无线网络所面临的另一个挑战。考虑到下一代无线网络,尤其是5G网络内急剧增加的IP业务量,我们在基站间采用大容量WDM(wavelength division multiplexing)有线光交换,通过在网络端点配置多个可调谐激光发射器(tunable lasers)实现基站同时发送多组IP业务包至不同小区。根据下一代无线网络“大规模,多参数化”的特点,针对网络能效,本文提出了“无线—有线—无线”一体化的系统能效联合优化目标。定义网络最大允许时延为受限时延,则我们通过受限时延在有线和无线之间的分配以及有线与无线资源消耗之间的折衷两个方面提出了具有时延限制功能的高能效业务调度,路由及无线发送一体化算法。所提算法能根据系统时延界限,在联合优化无线及有线参数的同时提升一体化网络的系统能效。通过仿真,我们验证了所提算法的有效性。我们基于纳什非合作博弈论,几何规划,凸优化等方法,针对不同网络优化目标及场景提供了有效的解决方案。同时,所提算法均符合下一代无线网络框架,也同样适用于其他无线网络的干扰管理和时延受限下的高能效传输问题。
[Abstract]:With the rapid development of the next generation of wireless communication services based on IP (Internet Protocol), the capacity of the wireless network and the backbone network service scheduling and routing between the base stations are facing new challenges. With the close attention of the academia and the industry, a series of funds are made to the wireless physical layer in the 4G (4th generation) network. Source integration to improve resource utilization, such as OFDMA (orthogonal frequency division multiple access), base station collaborative multipoint Technology (Co MP, coordinated multi-point), CA (carrier), etc.. Through resource integration, the network capacity can be greatly enhanced and the cell edge coverage is enhanced, but at the same time, interference management is also brought to the same time. Therefore, the interference coordination technology in the next generation wireless network presents new content, such as the formation of base station collaboration cluster and user scheduling in the multi-point collaboration of base station, power distribution of base station based on OFDMA. On the other hand, the rapidly increasing mobile IP service, especially the 5G (5th generation) wireless network physical layer, is put forward. Using mmwave, massive-MIMO (massive multiple-input multiple-output), ultra dense cell and other schemes to further support the wireless data rate of more than 1000 times higher than 4G, from the perspective of improving resource utilization and system energy efficiency, how to compromise between various "large-scale" resource dissipation and achieve time between base stations The service scheduling, routing and wireless transmission of extended limited functions are another challenge for next-generation wireless networks. Aiming at the problems mentioned above, this paper first studies the interference situation and proposes a practical interference coordination scheme based on the multi-point collaboration of base stations in the next generation wireless network. The actual transmission ability of the physical layer is further proposed from the point of view of the system energy efficiency in the wireless network. The proposed algorithm has the characteristics that can bind the maximum transmission delay (worst-case delay) of the data stream (delay). From the angle of wireless network interference management, we propose a set of distributed base station multipoint. The adaptive dynamic formation mechanism of the cooperative cluster has a strong adaptive ability to the dynamic change of the wireless communication environment. At the same time, the distributed decision mechanism is simple and extensible. After the base station cluster and the user scheduling result, we implement the classical WF (water-filling) power allocation algorithm in all the orthogonal frequency bands in the community. The power allocation strategy space to the base station in each band. Then, we use the Nash non cooperative power game method to model the power allocation in the base station collaboration cluster. Through the Nash game theory, we prove that the proposed game model has the unique Nash equilibrium point (NE, Nash Equilibrium), and the search for the only NE point search is given. The performance simulation shows that the proposed base station collaboration scheme can greatly improve the system capacity, and its distributed design is both simple and extensible, so it has a strong engineering feasibility. Although the heuristic algorithm has high engineering value, the transmission scheduling and scheduling of the maximum system capacity are realized from the theoretical point of view. The power allocation joint optimal solution is of great significance for evaluating the performance of the heuristic algorithm. However, the optimization problem is a hybrid optimization problem. In the existing literature, the two step heuristic solution of the user scheduling power allocation is almost all. In this paper, a novel PFM mechanism (power-fractionizing mechanism) is proposed to be discrete. The transmission scheduling of variables and the optimization of the power distribution of continuous variables are unified into the SP (SIGNOMIAL programming) optimization model. Based on the geometric programming optimization (GP, geometric programming), the local optimal solution of this kind of joint optimization problem is given. And from the experience of the GP algorithm, the local optimal solution is often the global optimal solution. The performance simulation results also verify the superiority and performance stability of the optimization algorithm than the heuristic algorithm. On the basis of the physical layer resources integration, how can the next generation wireless network, especially the "large-scale" resource consumption in the 5G, and the time delay between the lines and the wireless transmission are higher. Energy efficiency tradeoff is another challenge for a new generation of wireless networks. Considering the rapidly increasing IP traffic in the next generation wireless network, especially in the 5G network, we use a large capacity WDM (wavelength division multiplexing) wired optical switching between the base stations and configure multiple tunable laser transmitters (tunable LA) through the network endpoints. SERS) transmit multiple groups of IP packets to different communities at the same time. According to the characteristics of "large-scale, multi parameterized" in the next generation of wireless networks, in view of the network energy efficiency, this paper proposes a joint optimization target of "wireless to wired wireless" integrated system energy efficiency. Two aspects of the tradeoff between wired and wireless distribution and the tradeoff between wired and wireless resource consumption are proposed in this paper. A high energy efficient service scheduling, routing and wireless transmission integration algorithm with time delay limitation is proposed. The proposed algorithm can improve the integration of wireless and wired parameters at the same time according to the time delay limit of the system. Through simulation, we verify the effectiveness of the proposed algorithm. Based on the Nash non cooperative game theory, geometric programming, convex optimization and other methods, we provide an effective solution for different network optimization targets and scenes. At the same time, the proposed algorithms are all in line with the next generation wireless network framework, and are also applicable to the others. Interference management of line network and high efficiency transmission under delay constraint.

【学位授予单位】：电子科技大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TN929.5

【相似文献】