无线传感网络中拥塞控制与路由的跨层设计:分布式牛顿法

无线传感网络应用广泛, 其性能与路由选择和拥塞控制密切相关. 致力于拥塞控制与多径路由的跨层优化, 以实现在链路容量受限和节点能量受限情况下的无线传感网络效用最大化. 针对对偶次梯度算法具有收敛速度慢与信息交互量大等缺陷, 设计了具有二阶收敛性能的分布式牛顿算法来实现网络效用最大化. 通过矩阵分裂技术, 实现了只需单跳信息交互的牛顿对偶方向的分布式求解方法. 仿真结果表明, 分布式牛顿算法的收敛性能显著优于对偶次梯度算法.     

快速在线分布式对偶平均优化算法

为提高分布式在线优化算法的收敛速度,对底层网络拓扑依次添边,提出一种快速的一阶分布式在线对偶平均优化（FODD）算法。首先,对于分布式在线优化问题,运用添边方法使所选的边与网络模型快速混合,进而建立数学模型并设计FODD算法对其进行优化求解。其次,揭示了网络拓扑和在线分布式对偶平均收敛速度之间的关系,通过提高底层拓扑网络的代数连通度改进了Regret界,将在线分布式对偶平均（ODDA）算法从静态网络拓展到时变网络拓扑上,并证明了FODD算法的收敛性,同时解析地给出了收敛速度。最后的数值仿真表明:和ODDA算法相比,所提出的FODD算法具有更快的收敛速度。     

无线传感网中混合业务下的触发式流控制算法

无线传感器网络中,节点所具有的能量和通信能力等都十分有限,如何设计有效的协议及算法,利用有限的资源高效地完成诸多任务,成为无线传感器网络设计所面临的一大挑战.考虑接收容量模型,研究了无线传感器网络在节点接收容量和能量联合受限情况下,面向混合业务时的效用公平流控制问题,并针对传统对偶分解算法存在着收敛速度慢、步长不易调节、通信负荷大等缺陷,进一步提出了基于事件触发的分布式求解算法.理论分析与仿真验证均表明:使用事件触发算法时,传感节点的平均广播周期比使用对偶分解算法时大很多,大幅度降低了无线传感器网络节点间的通信量,减少了网络的通信开销.仿真结果显示:与对偶分解算法相比,分布式事件触发算法具有收敛速度快、对网络规模扩展的适应性强等优势;与传统的速率公平流控制机制相比,所提的效用公平流控制模型能够更加适应弹性与非弹性业务共存的网络场景.     

无线传感网络移动节点位置并行微粒群优化策略

网络节点位置优化是无线传感网络研究的核心问题之一.无线传感网络通常由固定节点和少量移动节点构成,传统的虚拟力导向算法无法解决固定节点对移动节点优化的约束.该文针对这一问题,提出了基于并行微粒群算法的优化策略.微粒群算法具有适于解决连续空间多维函数优化问题、能快速收敛至全局最优解的特点.并行框架提高了算法的运行效率,降低了算法的运算复杂度,使算法能够满足无线传感网络的需求.通过并行微粒群算法搜索不同状态下无线传感节点的最优位置,使无线传感网络能够利用移动节点实现网络结构的动态重组,最大化网络覆盖范围,提高网络测量可靠性.实验证明,并行微粒群优化策略能快速有效地实现无线传感网络移动节点位置优化.     

基于虚拟力的LoRa浮标网络覆盖优化算法

无线传感网络（WSN）节点部署问题是目前无线传感网络应用研究的关键点。针对传统网络节点部署存在收敛速度慢、全局优化性能不强、感知角度受限的问题,提出一种虚拟力导向的全向感知覆盖算法（VFOPCA）。该算法在传统虚拟力算法的基础上提出热点区域与节点间的受力模型,并采用0/1圆盘覆盖模型,对网络节点部署进一步优化。实验仿真表明,虚拟力导向的全向感知覆盖算法能快速有效地实现网络节点全局优化部署,与VFA、DACQPSO等全向感知模型算法相比,该算法覆盖程度更好、收敛速度更快、能耗程度更低。     

高吞吐率和高公平性混合传感网络

人们对传感网络吞吐率和公平性的要求越来越高,但是利用现有无线传感网络技术改善其传输性能却是非常困难的,因此基于现实工程中存在的一类特殊应用场景,设计了利用混合传感网络来改善传统无线传感网络低吞吐率以及低公平性的方法。首先针对固定传感网络论证了其最优吞吐率分配机制,其次针对网络布线问题设计了贪婪算法、K-自增聚类算法和混合算法3种启发式算法。仿真结果表明,混合算法相对于其他两种算法而言,网络最小节点吞吐率至少提高了75%,具有最优的算法性能,可以显著改善传感网络的性能。     

基于测量矩阵优化的分布式压缩估计算法

为更好地解决无线传感网分布式测量中有效数据估计问题,提出一种新的分布式压缩估计算法。通过在一个压缩维度上完成未知参数变量的分布式估计,并采用自适应随机梯度递归方法更新测量矩阵,将分布式压缩估计与测量矩阵优化相结合,实现收敛速度及估计误差精度的最优化。仿真结果表明,与d NLMS、DCE算法相比,该算法具有更快的收敛速度及更高的估计误差精度。     

基于改进蚁群算法的WSN移动代理路由算法研究

提出了基于改进蚁群算法的无线传感器网络移动代理路由算法,在改进算法中引入了传感节点的剩余能量值、数据处理能力等新的启发因素,从而均衡了网络负载,降低了网络能耗和延时;状态转换规则的改进和自适应全局信息素更新策略的采用克服了基本蚁群算法的不足。仿真实验表明,提出的算法在全局性和收敛速度上均优于其他传统算法。     

基于多粒子群算法的WSNs覆盖优化策略研究*

为了改善无线传感网络的性能,提高网络的覆盖率,在粒子进化的多粒子群算法的基础上,提出了一种无线传感网络覆盖的优化策略。该策略通过多个粒子群彼此独立地搜索解空间, 提高了算法的寻优能力,有效地避免了基本粒子群算法容易出现的“早熟”问题,提高了算法的稳定性。仿真实验表明,与基本粒子群算法、传统遗传算法和新量子遗传算法的优化效果相比较,其覆盖率分别提高了8.39％、3.07％和0.75％;收敛速度提高了25.3％、23.8％和23.8％。因此粒子进化的多粒子群优化策略具有比这三种算法更好的覆盖优化效果。     

基于分布式烟花算法的WSN布局优化方法

无线传感器网络具有广泛的应用,然而如何有效部署无线传感器节点,提高节点利用率和网络覆盖率,仍是一个亟待解决的问题。针对传统无线传感器网络部署方法存在节点冗余率高、覆盖率低等问题,以网络覆盖率为优化目标,将烟花算法良好的结果搜索能力和分布式高效的计算速度相结合,实现对网络覆盖率优化模型的高效求解。实验表明,该算法相比于普通的烟花算法具有更好的计算结果和更快的收敛速度。     

Utility-based scheduling in wireless multi-hop networks over non-deterministic fading channels

In this paper, an innovative scheduling scheme is proposed for interference-limited wireless multi-hop networks with non-deterministic fading channels. The scheduling problem is considered as a network utility maximization (NUM) problem subject to link rate constraints. By jointly taking into account of the link scheduling and the statistical variations of signal and interference power, the convex sets for the NUM are derived. Two types of non-deterministic fading channels (i.e., Rayleigh fading channel and Ricean fading channel) are characterized into our NUM models as examples. To solve the convex optimization problem, the subgradient projection method based on dual decomposition is employed. Then, a heuristic algorithm is designed for the TDM mode wireless multi-hop networks by minimizing the discrepancy between the expected network cost and the optimal one in each timeslot. At last, the source–destination session rate and network utility are evaluated in a dedicated wireless multi-hop network scenario. The numerical results demonstrate that the session rates convergence and the network utility is improved by our proposed scheme.     

Rate-lifetime tradeoff for reliable communication in wireless sensor networks

The network lifetime and application performance are two fundamental, yet conflicting, design objectives in wireless sensor networks. There is an intrinsic tradeoff between network lifetime maximization and application performance maximization, the latter being often correlated to the rate at which the application can send its data reliably in sensor networks. In this paper we study this tradeoff by investigating the interactions between the network lifetime maximization problem and the rate allocation problem with a reliable data delivery requirement. Severe bias on the allocated rates of some sensor nodes may exist if only the total throughput of the sensor network is maximized, hence we enforce fairness on source rates of sensor nodes by invoking the network utility maximization (NUM) framework. To guarantee reliable communication, we adopt the hop-by-hop retransmission scheme. We formulate the network lifetime maximization and fair rate allocation both as constrained maximization problems. We characterize the tradeoff between them, give the optimality condition, and derive a partially distributed algorithm to solve the problem. Furthermore, we propose an approximation of the tradeoff problem using NUM framework, and derive a fully distributed algorithm to solve the problem.     

Jointly optimal congestion control, channel allocation and power control in multi-channel wireless multihop networks

In this paper, to increase end-to-end throughput and energy efficiency of the multi-channel wireless multihop networks, a framework of jointly optimize congestion control in the transport layer, channel allocation in the data link layer and power control in the physical layer is proposed. It models the network by a generalized network utility maximization (NUM) problem with elastic link data rate constraints. Through binary linearization and log-transformation, and after relaxing the binary constraints on channel allocation matrix, the NUM problem becomes a convex optimization problem, which can be solved by the gateway centralized through branch and bound algorithm with exponential time complexity. Then, a partially distributed near-optimal jointly congestion control, channel allocation and power control (DCCCAPC) algorithm based on Lagrangian dual decomposition technique is proposed. Performance is assessed through simulations in terms of network utility, energy efficiency and fairness index. Convergence of both centralized and distributed algorithms is proved through theoretic analysis and simulations. As the available network resources increase, the performance gain on network utility increases.     

A joint maximum likelihood algorithm for simultaneous track correlation and sensorbias estimation

In this study a joint maximum likelihood （JML） algorithm was developed to solve problems re- garding interdependent and contradictory relationships between track correlation and sensor bias estimation in multi-sensor information fusion systems. First, the relationships between track correlation and sensor bias estimation of a multi-sensor system were analyzed. Then, based on these relationships, a JML function of the track correlation and sensor bias estimation was developed, while an iterative two-step optimization procedure was adopted to solve the JML function. In addition, transformation of sensor bias from Cartesian coordinates to polar coordinates and a complete design of track quality and ambiguity processing were provided. Finally, several Monte Carlo simulations were built to test the effect of target density and different sensor bias in the JML algorithm. Simulation results showed that the JML algorithm presented in this paper had a higher correct correlation rate and more accurate sensor bias estimation than traditional methods, demonstrating that the JML algorithm had good performance.     

Alternative Distributed Algorithms for Network Utility Maximization: Framework and Applications

Network utility maximization (NUM) problem formulations provide an important approach to conduct network resource allocation and to view layering as optimization decomposition. In the existing literature, distributed implementations are typically achieved by means of the so-called dual decomposition technique. However, the span of decomposition possibilities includes many other elements that, thus far, have not been fully exploited, such as the use of the primal decomposition technique, the versatile introduction of auxiliary variables, and the potential of multilevel decompositions. This paper presents a systematic framework to exploit alternative decomposition structures as a way to obtain different distributed algorithms, each with a different tradeoff among convergence speed, message passing amount and asymmetry, and distributed computation architecture. Several specific applications are considered to illustrate the proposed framework, including resource-constrained and direct-control rate allocation, and rate allocation among QoS classes with multipath routing. For each of these applications, the associated generalized NUM formulation is first presented, followed by the development of novel alternative decompositions and numerical experiments on the resulting new distributed algorithms. A systematic enumeration and comparison of alternative vertical decompositions in the future will help complete a mathematical theory of network architectures.     

A Rate Allocation Framework for Multi-Class Services in Software-Defined Networks

Software defined networking (SDN) is a network architecture with a programmable control plane (e.g., controllers) and simple data plane (e.g., forwarders). One of the popular SDN protocols/standards is OpenFlow, for which researchers have recently proposed some quality-of-service (QoS) supports. However, the proposals for rate allocation have some limitations in network scalability and multi-class services’ supports. In the literature, rate allocation formulations are commonly based on the framework of network utility maximization (NUM). Nevertheless, multi-class services are rarely considered in that framework since they make the formulated NUM become nonconvex and prevent its subgradient-based algorithm from converging. In this paper, we propose a scalable QoS rate allocation framework for OpenFlow in which multi-class services are considered. The convergence issue in the algorithm of our NUM-based framework is resolved by an admission control scheme. The network scalability is improved by our decentralized algorithms that can run on multiple parallel controllers. Extensive simulation and emulation results are provided to evaluate the performance of our method.     

编码分组网络的效用最大化研究

对基于网络编码方案的分组网络(即编码分组网络)的效用最大化问题进行研究。利用网络编码和网络流的对应关系以及组播树分解方法提出单通话编码分组网络效用最大化模型。基于对偶分解理论推导出解决单通话编码分组网络效用最大化问题的分布式次梯度投影算法,找到一个有效的Lipschiz常数从而得到算法收敛的充分条件。通过仿真验证了该算法的正确性。     

Primal and dual-based algorithms for sensing range adjustment in WSNs

Coverage and tracking of multiple targets, are viewed as important challenges in WSNs, mainly aimed for future ubiquitous and pervasive applications. Target coverage in WSNs with large numbers of sensor nodes and targets, and with a predefined placement of sensors, may be conducted through adjusting the sensing range and considering the energy consumption related to this operation. In this paper, we encounter the problem of multiple target coverage in WSNs by determining the sensing range of each sensor node to maximize the total utility of the network. We solve this Network Utility Maximization (NUM) problem via two approaches, primal and dual decompositions, which result in iterative distributed price-based algorithms. Convergence of sensing ranges to optimal values is proved by means of stability analysis and simulation experiments. Simulation results show convergence to optimal values in few iterations, with near optimal values for the total objective function and energy consumption of nodes. These results show scalability of our algorithm, in terms of the number of iterations needed for convergence when compared with the other two methods. Furthermore, the distributed algorithm based on dual decomposition is used to cover efficiently moving targets in consecutive time intervals.     

Distributed rate control,routing, and energy management in dynamic rechargeable sensor networks

This paper studies the network utility maximization (NUM) problem in dynamic-routing rechargeable sensor networks (RSNs), where rate control, routing, and energy management need to be jointly optimized. This problem is very challenging since the flow constraint is spatially coupled and the energy constraint is spatiotemporally coupled (energy causality). Existing works either do not fully consider the two coupled constraints together, or heuristically remove the temporally-coupled part, both of which are not practical, and may degrade network performance. In this paper, we attempt to jointly optimize rate control, routing, and energy management by carefully tackling the flow and energy constraints. To this end, we first decouple the original problem equivalently into separable subproblems by means of dual decomposition. Then, we propose a distributed algorithm, which can converge to the globally optimal solution. Numerical results based on real solar data are presented to evaluate the optimality and scalability of the proposed algorithm.