基于模糊神经网络的异构无线网络接入选择算法

针对现有异构无线网络基于模糊逻辑及神经网络的接入选择方法未能合理考虑网络负载状况的问题,提出一种基于RBF(径向基函数)模糊神经网络的接入选择方法.该方法以可接入网络的接入阻塞率相等为模糊神经网络参数强化学习的目标,对网络负载程度具有很好的动态适应性,实现了智能化的接入判决.仿真结果表明,该方法能有效均衡异构无线网络间的负载,保障实时与非实时业务的QoS,并且相对于负载均衡算法(MLB算法)降低了网络的接入阻塞率.     

基于遗传聚类的无线传感器负载均衡路由算法

通常的无线传感器分簇网络存在节点负载不均衡的问题。为均衡各节点能量消耗,延长网络生存周期,将K均值算法与遗传算法相结合,提出一种负载均衡的无线传感器网络路由算法,算法利用遗传算法的全局寻优能力以克服传统K均值算法的局部性和对初始中心的敏感性,实现了传感器网络节点自适应成簇与各节点负载均衡。仿真实验表明,该算法显著延长了网络寿命,相对于其他分簇路由算法,其网络生存时间延长了约43%。     

基于遗传聚类的无线传感器负载均衡路由算法

通常的无线传感器分簇网络存在节点负载不均衡的问题。为均衡各节点能量消耗,延长网络生存周期,将K均值算法与遗传算法相结合,提出一种负载均衡的无线传感器网络路由算法,算法利用遗传算法的全局寻优能力以克服传统K均值算法的局部性和对初始中心的敏感性,实现了传感器网络节点自适应成簇与各节点负载均衡。仿真实验表明,该算法显著延长了网络寿命,相对于其他分簇路由算法,其网络生存时间延长了约43%。     

基于量子遗传算法的无线视频传感网络优化覆盖算法

针对智慧城市无线视频传感网络建设需要,提出一种基于量子遗传算法的网络优化覆盖算法。算法面向复杂的监视区场景,监视区中存在形状各异的障碍物,各区域的重要程度不同。以二维离散网格模型描述监视区场景,用编码描述矩阵表示监视区域,用七元组描述有向无线视频传感器。通过严格的数学推导得出了问题的数学规划模型。优化覆盖算法由IntialDeployment算法和OptimizedDeployment算法2部分组成,以获得最大有效覆盖率的网络部署方案为求解目标。采用量子遗传算法搜索解空间,通过合理编码染色体,优化量子旋转门参数,使算法的运算速度快,收敛性好。引入理想覆盖率和理想加权覆盖率2个极限值,采用相对比较法评判算法优劣。仿真实验和数据分析表明,算法获得的方案能很好地逼近理想极限值。在传感器节点数给定的情况下,算法能获得最大的覆盖率。     

基于改进的遗传算法无线通信网频率指配问题研究

通过对无线通信频率指配问题的分析,结合遗传算法在频率指配领域的应用,提出了一种启发式的指配方法。该方法通过改进选择方式,自适应地调整交叉、变异概率来指配信道分配。仿真分析证明,该算法科学可行,有效地避免陷入局优解,加快了种群进化速度,减少了迭代次数,较快收敛到最优解。     

An energy-aware routing protocol for wireless sensor network based on genetic algorithm

Energy saving and effective utilization are an essential issue for wireless sensor network. Most previous cluster based routing protocols only care the relationship of cluster heads and sensor nodes but ignore the huge difference costs between them. In this paper, we present a routing protocol based on genetic algorithm for a middle layer oriented network in which the network consists of several stations that are responsible for receiving data and forwarding the data to the sink. The amount of stations should be not too many and not too few. Both cases will cause either too much construction cost or extra transmission energy consumption. We implement five methods to compare the performance and test the stability of our presented methods. Experimental results demonstrate that our proposed scheme reduces the amount of stations by 36.8 and 20% compared with FF and HL in 100-node network. Furthermore, three methods are introduced to improve our proposed scheme for effective cope with the expansion of network scale problem.     

基于遗传算法的无线传感器网络节点优化方法研究

随着科技的不断发展,传感器网络得到了极大的发展,其拥有极为广阔的应用前景,已被应用于智能家居、空间探索、军事、健康护理以及环境监测等领域。文章就其测控系统所具有的特殊性,提出了一种可以达到节点连通约束来对节点数量加以减少的新模型,同时利用遗传算法对其加以优化计算。要想使节点收敛速度大幅提升,文章将采取以二分法为基础的遗传算法进行研究。     

Relay algorithm based on network coding in wireless local network

The network coding is a new technology in the field of information in 21st century. It could enhance the network throughput and save the energy consumption, and is mainly based on the single transmission rate. However, with the development of wireless network and equipment, wireless local network MAC protocols have already supported the multi-rate transmission. This paper investigates the optimal relay selection problem based on network coding. Firstly, the problem is formulated as an optimization problem. Moreover, a relay algorithm based on network coding is proposed and the transmission time gain of our algorithm over the traditional relay algorithm is analyzed. Lastly, we compare total transmission time and the energy consumption of our proposed algorithm, Network Coding with Relay Assistance (NCRA), Transmission Request (TR), and the Direct Transmission (DT) without relay algorithm by adopting IEEE 802.11b. The simulation results demonstrate that our algorithm that improves the coding opportunity by the cooperation of the relay nodes leads to the transmission time decrease of up to 17% over the traditional relay algorithms.     

基于遗传算法的蜂窝网络接入信道动态分配方案的设计

蜂窝网络中,使用信道复用技术能够在很大程度上提高移动通信的容量和质量。但是,信道复用过程中会产生不同程度的电子干扰问题,使得信道复用的作用大打折扣。传统的进行信道优化的方法很多,但是效果都不是很理想。这里利用遗传算法进行蜂窝网络接入信道的动态分配,主要通过对集中常出现的信道干扰问题加上一些约束条件,这样建立的数学模型能够获得一最小干扰信号的信道分配方案,在很大程度上避免了移动用户之间的干扰问题。     

Dynamic clustering approach based on wireless sensor networks genetic algorithm for IoT applications

Energy is vital parameter for communication in Internet of Things (IoT) applications via Wireless Sensor Networks (WSN). Genetic algorithms with dynamic clustering approach are supposed to be very effective technique in conserving energy during the process of network planning and designing for IoT. Dynamic clustering recognizes the cluster head (CH) with higher energy for the data transmission in the network. In this paper, various applications, like smart transportation, smart grid, and smart cities, are discussed to establish that implementation of dynamic clustering computing-based IoT can support real-world applications in an efficient way. In the proposed approach, the dynamic clustering-based methodology and frame relay nodes (RN) are improved to elect the most preferred sensor node (SN) amidst the nodes in cluster. For this purpose, a Genetic Analysis approach is used. The simulations demonstrate that the proposed technique overcomes the dynamic clustering relay node (DCRN) clustering algorithm in terms of slot utilization, throughput and standard deviation in data transmission.

     

基于遗传算法的非监督层次分类方法研究

快速准确地确定单个样本的所属类别以及总体样本类别数是解决非监督模式识别的前提,然而它们的确定通常是非常困难的.通过研究基于遗传算法的相似性度量最优分类算法以及最优分类数确定算法,提高非监督识别的准确性,并将所研究的算法应用到飞机识别当中.实验结果表明,本算法可以进行最优分类及分类数的确定.     

Designing a mobile broadband wireless access network

In this paper, a new OFDM-based air interface technology for a mobile broadband wireless system is described. The technology leverages the standard Internet protocol (IP) network elements to build the system and deploys a new air interface technology based on OFDMA. Cross-layer optimization played a major role in the design where the choices made in the physical, MAC, and link layers are also driven by the goal of extending the Internet to the wireless space. A major physical layer benefit of this air interface comes from the orthogonality property that the results in the elimination of in-cell interference are averaged and a worst-case interferer does not limit the system performance. The physical layer features not only result in high capacity but also provide very fine granularity of allocating air link resources, which improves the MAC and link-layer efficiency. The MAC and link layer provide contention-free, fast control channels between the RAR and the WTs. These channels are used to ferry a variety of signaling such as assignments of traffic channel, acknowledgements, channel quality, and traffic request reports. This holistic approach allows for a scheduler that could not only achieve high spectral efficiency but also allow for a fine control over QoS attributes such as latency, reliability, and service differentiation.     

基于关联规则引导遗传算法的高可靠性无线网络拓扑设计

无线网络中的节点与路径故障会产生惩罚性网络成本,该成本是无线网络的一个重要性能指标,对此提出了一种基于关联规则引导遗传算法的高可靠性无线网络拓扑设计算法。首先,采用Monte Carlo模拟器将网络模拟为图结构;然后,采用Apriori算法提取模拟器数据的关联规则;最后,利用提取的关联规则引导遗传算法的变异与交叉操作,搜索最优的网络拓扑结构。仿真实验结果表明,对于多个网络规模,该算法均可获得较好的网络性能与收敛速度,具有较好的实用性。     

A new wireless network medium access protocol based on cooperation

In this paper, we propose a new media access protocol for wireless networks, that due to its ability to resolve collisions can achieve high throughput. We view the wireless network as a spatially distributed antenna with antenna elements linked via the wireless channel. When there is a collision, the collided packets are saved in a buffer. In the slots following the collision, a set of nodes designated as nonregenerative relays retransmit the signal that they received during the collision slot. By processing the originally collided packets and the signals forwarded by the relays, the destination node can recover the original packets. The proposed scheme maintains the benefits of ALOHA systems, i.e., needs no scheduling overhead and is suitable for bursty sources, such as multimedia sources. It also offers the benefits of multi-antenna systems, i.e., spatial diversity while employing a single transmit/receive antenna at each node. Spatial diversity enables it to be robust to the wireless channel. The proposed approach achieves higher throughput and energy savings than existing techniques that allow for multiple packet reception.     

基于改进遗传算法的无线网络覆盖算法

考虑到传统的遗传算法在对无线传感网络覆盖进行优化时,存在起始阶段计算速度快,后期局部寻找最优解能力弱,不能充分使用系统反馈路径信息,使得算法会因冗余迭代而导致陷入局部最优解,影响优化效率和覆盖率等问题,将蚁群算法融合到遗传算法中,对遗传算法进行改进。通过不同覆盖范围和节点的三个实例进行优化效果分析,可知在小面积的覆盖范围内,以及节点个数较少时,该文研究的改进方法与传统优化方法的覆盖率和完成时间差别不大,但是随着覆盖范围的增大,节点个数的增加,该文研究的改进方法完成时间明显缩短,覆盖率明显增大,相比传统优化方法具有更大的优势。     

基于ACS的无线传感器网络区分服务路由算法

针对无线传感器网络中数据传输的不同要求,将QoS分为3类,根据无线链路的特点提供区分服务。利用博弈论分析了数据传输在延迟、可靠性与网络能量开销之间的关系,基于改进的蚁群优化算法ACS(ant colony system),设计了区分服务路由算法ADSGR(ant colony system based differentiated service and game-theory routing),依据不同QoS要求,选择适当的路由,提高网络的整体性能和资源利用率。实验结果表明,与现有算法相比,该算法在数据传输的延迟、可靠性和能量开销上具有更好的性能。     

基于数据融合的无线传感器网络路由算法

在分簇协议LEACH和链状协议PEGASIS的基础上,提出一种新的基于数据融合的分簇路由算法.簇首节点采用多跳方式传输数据,并根据周围节点的密集程度构造不同大小的簇;簇内节点计算上行和下行节点构造数据融合树,采用时分复用调度算法进行多跳路由.NS2仿真结果表明该路由算法均衡了各个节点的能量消耗,延长了网络存活时间,并降低了网络延迟.     

基于SDN的工业无线网络能量高效型路由算法

在智能制造背景下,为了提高无线网络的灵活性和能量使用效率,提出基于软件定义网络(SDN)的工业无线网络能量高效路由算法(SDERA)。基于SDN数据平面和控制平面分离的思想,构建了新型的工业无线网络模型,控制器根据应用需求、节点的剩余能量和负载等因素,计算全局最优路由策略并分发给交换节点,交换节点按照控制器的决策在不同的逻辑子网中转发数据包。当节点和链路质量信息以及应用需求发生变化时,控制器可以动态地调整路由策略,提高能量使用效率和网络性能。仿真实验结果表明,该路由算法有效地延长了网络的生存时间,提高了网络的能量使用效率。     

基于扇形分簇的无线传感器网络路由算法

无线传感网络中低功耗自适应聚类分簇(LEACH)路由算法等概率选取簇首节点,容易导致整个网络节点能量损耗出现极端化,减少网络生存时间。为此,提出一种针对簇首节点选取和分簇的改进LEACH算法。该算法把整个网络区域分为四个扇形区域,在每个区域内独立进行分簇路由;然后基站根据节点剩余能量和与基站的距离进行簇首节点选择,节点根据簇首节点和基站接收信号强度选择路由方式,以均衡网络能量消耗。仿真结果表明,改进LEACH算法的网络寿命是原有LEACH算法的150%,数据吞吐量提升了3倍。     

基于二次质心的无线传感器网络定位算法

节点的定位是无线传感器网络中的一种重要技术。提出了一种新的无线传感器网络定位算法——基于二次质心算法的定位算法,与以往的基于三边测量的加权质心方法不同,该算法改进了对未知节点位置的估算方法,一定程度上避免了因多次估算质心而产生的累积误差,提高了定位精度。仿真表明,该算法的定位精度较之前的三边测量方法提高了约19%。