基于卷积神经网络的新奇检测技术在结构损伤识别中的应用

针对新奇检测难以同时识别结构损伤时刻和损伤位置的问题,提出在新奇检测中引入卷积神经网络以实现损伤时刻和损伤位置的一次性确定。首先,采用小波包技术处理结构响应得到小波包能量,并将相邻测点对应频带的能量比作为新奇检测模型的特征向量;然后,以结构健康时的特征向量作为训练数据,建立健康模式下的基于卷积神经网络的新奇检测模型;接着,将结构实时输出的特征向量输入到新奇检测模型,所得输出与健康状态的输出进行对比,并将输出和输入的欧氏距离作为新奇指标;最后,根据新奇指标的变化识别结构损伤时刻和损伤位置。数值模拟和实验室试验验证了该方法的有效性。     

基于帕累托最优效率模型的轨道交通沿线公交线路运行效率评价

通过建立公交运行效率评价模型,确定轨道交通沿线待优化线路,为两网融合奠定理论基础。研究首先通过定性定量法筛选出与轨道交通关联度大且有研究意义公交线路,其次以公交线网自身合理性、两网之间协同关系等数据构建评价体系,最后采用帕累托最优交叉效率模型,求解沿线公交线路运行效率。以南昌市1号线为例,筛选出沿线25条公交线路并进行实证分析。结果表明：帕累托最优交叉效率模型的评分更容易被所有的决策单元（Decision-Making Units,DUM s）接受;数据分析得出南昌公交线网自身合理性较好,但在两网协同关系方面,可优化线路走向、调整线路重复系数等加强其协同性。最后基于两者的帕累托交叉效率均值,将25条公交线路划分为4种案例,为公共交通系统的管理者和决策者提供了更实用和有价值的观点。     

基于YOLOv5s-FCS的钢材表面缺陷检测

针对传统钢材表面缺陷检测方法易出现误检、漏检和部分缺陷种类检测精度低等问题,本文设计了一种钢材表面缺陷网络YOLOv5s-FCS。首先本文引用了FReLU激活函数构建了卷积模块CBF,有效增强了网络的空间解析能力,优化了网络检测精度;其次,本文将坐标注意力机制嵌入到网络的neck部分来增强网络特征融合的能力,从而使网络能够提取更加丰富的特征信息;最后,将YOLOv5s的损失函数替换为SIoU loss,提高了预测框的回归精度。通过在NEU-DET数据集上进行消融实验、可视化对比实验,结果表明,YOLOv5s-FCS网络的mAP值达到了0.747,相较于原YOLOv5s网络提高了8.3%,相较于YOLOv3网络提高了11.8%,相较于YOLOXs网络提高了4.2%,相较于YOLOv6s提高了1.4%,验证了该方法的可行性、有效性。     

一种面向在轨服务卫星地面应用系统的软件开发方法

为解决现有在轨服务卫星地面应用系统瀑布式软件开发模型灵活性差,难以适应用户需求变化等不足,通过利用敏捷开发模型的思路,设计了一种迭代式在轨服务卫星地面应用系统软件开发方法,将用户纳入到软件开发过程中,在各个研制阶段持续与用户进行需求迭代,完成业务软件的功能升级。结果表明：该方法可提升业务软件功能与用户需求的匹配程度,并为后续在轨服务卫星地面应用系统的建设提供参考。     

石榴枯萎病的生防木霉菌株的筛选及鉴定

目的 石榴枯萎病是由甘薯长喙壳菌引起,对石榴有严重危害的土传病害。利用拮抗微生物抑制病原菌的繁殖是石榴枯萎病绿色防控的有效手段。通过筛选具有对石榴枯萎病原菌产生拮抗效果的微生物,研究其抑制菌作用,以期为石榴枯萎病生物防控提供菌种资源。方法 从石榴的根际土壤中分离出真菌,根据产孢量、生长速率、拮抗系数及平板对峙结果进行筛选,对防效较好的真菌根据形态学和分子系统学进行鉴定,通过盆栽实验评估其对石榴枯萎病的防治效果。结果 能有效抑制石榴枯萎病菌菌丝的生长的菌株T9经鉴定为非洲哈茨木霉菌（Trichoderma afroharzianum）,盆栽防治效果最高可达71.35%。结论 该菌株展现出了对石榴枯萎病的生物防治效果,具有应用价值。     

热浸Zn-Ni-V合金镀层的氧化行为

研究了热浸纯Zn和Zn-0.05%Ni-0.05%V镀层在350℃下的恒温氧化行为。采用SEM、OM、EDS、XRD等方法分析了镀层氧化前后的表面、截面形貌及相组成。结果表明:Zn-0.05%Ni-0.05%V镀层可以有效地抑制Fe-Zn反应和ζ相层的超厚生长;350℃下氧化150h后,纯Zn和Zn-0.05%Ni-0.05%V镀层由ZnO和FeZn8.87两相组成;各镀层氧化动力学曲线主要为抛物线类型;在氧化过程中,纯Zn镀层形成大量的孔洞,氧化增重速率快,而Zn-0.05%Ni-0.05%V镀层组织较致密,具有更优的抗氧化性能     

Research on secure buyer-seller watermarking protocol

A new buyer-seller watermarking protocol is proposed by applying a double encryption method and a novel mechanism of embedding a buyer＇s watermark. The protocol can effectively prevent against collusion attacks and the man in the middle attack if the third party is not trusted. Also, based on the proposed scheme for the first-hand transaction, a new buyer-reseller watermarking protocol and a formal multi-party watermarking protocol are also proposed. The proposed buyer-resell watermarking protocol only needs the original seller to provide transfer certificate and encryption-decryption service to support the second-hand transaction, and the multi-party watermarking protocol with distributed certificate authorities can overcome the difficulty in the combination of multicast mechanism with multiple unique watermarks and allow a seller to multicast the watermarked digital contents and key transaction information to n buyers. Furthermore, the idea of zero knowledge proof is also applied into the proposed scheme to allow the seller to take an effective control on the task performed by the third party.     

Data fusion control and guidance of surface-to-air missile under the complex circumstance based on neural-net technology

Under the complicated electromagnetism circumstance, the model of data fusion control and guidance of surface-to-air missile weapon systems is established. Such ways and theories as Elman-NN, radar tracking and filter＇s data fusion net based on the group method for data-processing （GMRDF） are applied to constructing the model of data fusion. The highly reliable state estimation of the tracking targets and the improvement in accuracy of control and guidance are obtained. The purpose is optimization design of data fusion control and guidance of surface-to-air missile weapon systems and improving the fighting effectiveness of surface-to-air missile weapon systems.     

Bell state preparation based on switching between quantum system models

For the preparation of any target Bell state under continuous quantum measurement, this paper proposes a method which achieves the control objective by switching between two different models or by switching between two control channels under one model. Proper control Hamiltonians are selected for the two system models, a switching strategy between the two models is designed, and the stability of the whole switching system is proved in theory. For a given target Bell state, the effectiveness of the proposed switching control strategy between different models is illustrated through simulation experiments.     

Optimal bundling and pricing decisions for complementary products in a two-layer supply chain

This article explores bundling and pricing decisions for two complementary products in a two-layer supply chain consisting of a multi-product manufacturer and a retailer. We establish four different pricing models under cases of decentralized decision, while considering different portfolio-bundling strategies of the manufacturer and the retailer. A game-theoretical method is used to characterize the corresponding equilibrium outcomes in each scenario. By further analyzing and comparing the maximum profits of all four possible scenarios, optimal bundling and pricing decisions for the manufacturer and the retailer are obtained, respectively. Model extensions and numerical examples are enriched to highlight the factors affecting optimal decision-making. Finally, valuable and interesting managerial insights are summarized. Results show that the upstream manufacturer always profits more when he sells complementary products separately. However, the optimal bundling decision of the downstream retailer is jointly determined by product complementarity (as a major factor) and the difference of product profitability (as a secondary factor). Market power cannot exert an influence on both optimal bundling decisions, but it can partly affect pricing decisions.