空中异构战术网络IP通信架构及协议开销分析

空中异构战术无线网络由多种不同类型的无线数据链网络构成,IP通信技术可有效解决空中平台异构无线网络的互联互通问题。给出了空中异构战术无线网络模型,并根据不同类型的无线数据链网络特点,分析了端到端通信协议栈关系以及不同无线数据链网络间协议转换与适配方式。探讨了网络静态和动态IP地址分配方法以及网络的路由架构与寻址方式,提出了不同无线数据链网络间IP报文传输的打包方式。对传输的开销性能仿真对比分析表明,基于通用成帧协议的打包方式的协议开销性能较优。     

无线Ad Hoc网络中基于活跃节点数预测的时隙ALOHA算法

针对采用时隙ALOHA算法的无线Ad Hoc网络,考虑数据包到达的动态性与数据包传输的随机性,以最大化吞吐量为目标,同时满足数据队列稳定性,构建了关于接纳控制与竞争接入的随机优化问题。由于在时隙ALOHA算法中数据包传输的最优概率取决于无线Ad Hoc网络中数据队列非空的活跃节点数,提出了一种基于活跃节点数预测的时隙ALOHA算法。该算法要求无线Ad Hoc网络中的所有发送节点实时地侦听通信信道的忙闲状态,计算基于信道状态的活跃节点数条件期望,从而动态地预测无线Ad Hoc网络在不同时刻的活跃节点数,达到网络节点依据局部网络状态信息自适应地优化数据包传输概率的目的。仿真结果表明,所提算法能够有效估计无线Ad Hoc网络在每个时隙的活跃节点数,从而显著提升网络吞吐量并且降低数据包的平均排队时延。     

一种基于地理信息系统的异构网络多目标算法

地理信息系统（Geographic Information System,GIS）和多目标算法为异构网络基站部署问题带来新的突破。ArcGIS系列是GIS专业化的管理软件,采用ArcGIS软件对包含学校、医院等场景的中国某市密集城区进行建模分析,并且提出基于目标权重因子的多目标离散粒子群算法（Multi-objective Discrete Particle Swarm Optimization algorithm based on Target Weight Factor,MDPSO/TWF）,基于多个目标函数整体最优研究异构网络基站部署问题。仿真结果表明,所提的MDPSO/TWF算法具有可行性和先进性,可以综合考虑网络负载和网络能效多个优化目标,将业务需求与实际基站选址相结合,在满足业务需求的基础上尽可能达到节能的目的。     

改进的YOLOv3-Tiny网络在风机叶片损伤检测中的应用

为了解决YOLOv3-Tiny对无人机采集的风机叶片图像损伤检测精度不高的问题,提出一种基于深度学习的风机叶片图像损伤检测方法。首先提出一种跨越式特征联合网络结构,由卷积层和拼接层构成,将不同深度的特征信息进行融合再学习,提取目标多层级特征信息;其次引入Inception模块结构,其中4个平行通道的多个卷积核对输入的特征图进行组合和压缩,在减少网络的学习参数的同时更好地表征图像特征信息,提高小目标的检测精度。实验表明,改进后算法的检测精度提高了2.69%,在自制的数据集中mAP可以达到88.58%,并且模型的参数缩小了4倍。因此,改进的方法比传统的YOLOv3-Tiny网络具有更好的检测效果。研究结果可为基于图像的损伤检测和风机叶片损伤智能识别提供参考。     

基于多传输手段的自组网路由协议研究

为了实现多手段通信资源的全局共用与优化使用,针对通信节点多传输手段的自组网特点,提出了一种基于节点MPR的OLSR路由协议（N-OLSR）。首先对HELLO报文格式进行修改,加入Port字段,根据通信手段的不同,对Willingnes字段的值做出修改,用接口索引来代替接口IP地址,将链路带宽大的信道用大索引号表示。以此为基础,从全局出发制定路由策略,将通信节点作为MPR的选举对象,改变传统方法在每个接口上进行MPR选举的机制。最后,通过仿真实验来测试N-OLSR路由协议的性能。结果表明,改进的协议完成了MPR从接口到节点的转移,降低了源节点将一跳邻居节点选为MPR的频率,从而减少TC报文的产生以及洪泛;改进的路由协议能够减少控制开销,降低丢包率,加快网络收敛速度。研究结果可为多传输手段的移动自组网提供一种可靠的路由协议方案,并且信道带宽越小越能发挥其性能。     

连锁董事网络、技术高管与企业R&D投入

以我国2015-2017年技术密集型制造业上市公司数据为样本,基于网络中心度和弱联结程度两个层面,实证研究了连锁董事网络对我国企业R&D投入的影响,并探讨了技术高管对这种影响的调节作用。结果显示：网络中心度的提升对企业R&D投入具有显著的促进作用,该促进作用受到技术高管的正向调节;弱联结程度对企业R&D投入无显著独立影响,但与技术高管共同对企业R&D投入产生正向的联合影响。     

Spatial movement patterns among intra-destinations using social network analysis

To explore popularly visited tourist locations, travel movement patterns, and movement points, this study collected samples of 321 Chinese tourists and 337 Japanese tourists who were visiting major tourist destinations in Seoul and its vicinity in South Korea. Results of analyzing movement patterns showed that Japanese tourists tend to be clustered around popular attractions, whereas Chinese tourists tend to spread over a larger area of attractions. Some specific shopping and amusement attractions were the locations most popularly visited by both groups. The start points and end points in the two groups’ itineraries were dissimilar overall, even though their patterns were similar in regard to major preferred tourist attractions. Thus, the findings of this study have the potential to contribute to understanding spatial mobility in a tourism destination through tracking tourists’ movement patterns.     

Platform coopetition in the tourism industry: conflicts and tensions caused by the closure of Booking.com in Turkey

In March 2017, as a result of a precautionary interim decision made by the courts based on the case relating to ‘unfair competition practices’, Booking.com closed its platform for inbound tourism in Turkey and ceased operations. This was the first instance of an online booking platform halting access for inbound bookings for the domestic market. This platform-based relationship involves both cooperation and competition, and thus this paper identifies this relationship as platform coopetition. Based on the governance network theory and employing a case-study approach, the Booking.com case is examined from the viewpoint of tourism coopetition from two perspectives: industry-wide and agent-specific. The paper provides an analysis of these perspectives and the past and on-going process of Booking.com’s platform closure. The constitutive dynamics this case endangers the coopetitive environment of the tourism industry and thus creates destructive uncertainties, especially for small hotels. This analysis also reveals the issues in terms of political representation for digital service platform companies.     

A dynamic perspective: local interactions driving the spread of social networks

ABSTRACT

This study takes a micro standpoint to quantitatively verify the extent to which local interactions affect the expansion of dynamic social networks. By reviewing and synthesizing the theoretical literature in this article, we seek to establish a pragmatic analytical framework to capture the underlying functioning of local interactions. It then aims to produce findings on seven distinct marketing campaigns in which the real-world virtual communities offer intact time-series connections. In our considered cases, the firm-to-consumer interactivity plays a pivotal role in social media marketing while the customer-driven influence advances individual information transmission. The significant positive relationships demonstrate features of interactivity within subgroups including reciprocal edges and directed 3-cliques associated with the evolution of the virtual community by electronic word-of-mouth (WOM) of both firm-to-consumer and consumer-to-consumer interactions. More importantly, the attenuation of local interactions to a large extent leads to the cut-off point of the growth of the dynamic social network.     

Big data driven cycle time parallel prediction for production planning in wafer manufacturing

Cycle time forecasting (CTF) is one of the most crucial issues for production planning to keep high delivery reliability in semiconductor wafer fabrication systems (SWFS). This paper proposes a novel data-intensive cycle time (CT) prediction system with parallel computing to rapidly forecast the CT of wafer lots with large datasets. First, a density peak based radial basis function network (DP-RBFN) is designed to forecast the CT with the diverse and agglomerative CT data. Second, the network learning method based on a clustering technique is proposed to determine the density peak. Third, a parallel computing approach for network training is proposed in order to speed up the training process with large scaled CT data. Finally, an experiment with respect to SWFS is presented, which demonstrates that the proposed CTF system can not only speed up the training process of the model but also outperform the radial basis function network, the back-propagation-network and multivariate regression methodology based CTF methods in terms of the mean absolute deviation and standard deviation.