带组换装时间的单机调度问题

在某钢铁线材企业的实际调度问题的基础上,研究了一类带有组换装时间的单机调度问题.根据该调度问题的实际需求,以最小化作业的最大延迟为优化目标.由于该问题是NP难的,提出了一类启发式算法来求解该问题,并进一步通过引入问题的性质,提高算法的寻优性能,降低算法运行时间.该算法在随机产生的测试问题和企业的实际调度上均进行了测试,实验结果表明该启发式算法能在短时间内获取近优解.     

Environment-aware intelligent numerology control approach for 5G and beyond systems

5G and beyond (5GB) systems show a flexible characteristic to meet different user communications requirements in parallel. Multi-numerology waveform design is an important part of this human centric characteristic, however, it is necessary to make multi-numerology planning. In this paper, the integrated sensing and communications (ISAC) concept is used with machine learning (ML) techniques to develop an intelligent numerology control mechanism for 5GB. Thanks to the ISAC capabilities, an environment-aware system is designed, and several useful sensing information (USI) that affect the wireless channel characteristic is utilized. It is assumed that the USI is obtained through 5GB smart city networks. At that point, synthetic data generation is performed to form a new dataset with USI-based features. Additionally, a novel feature control algorithm is proposed in the paper. Lastly, simulation results are presented for different type of ML models including several ensemble methods.     

AI-Driven Wearable Mask-Inspired Self-Healing Sensor Array for Detection and Identification of Volatile Organic Compounds

Volatile organic compounds (VOCs) sensor arrays have garnered considerable attention due to their potential to provide real-time information for monitoring pollution levels and personal health associated concerning VOCs in the ambient environment. Here, an AI-driven wearable mask-inspired self-healing sensor array (MISSA), created using a simplified single-step stacking technique for detecting and identifying VOCs is presented. This wearable MISSA comprises three vertically placed breathable self-healing gas sensors (BSGS) with linear response behavior, consistent repeatability, and reliable self-healing abilities. For wearable and portable monitoring, the MISSA is combined with a flexible printed circuit board (FPCB) to produce a mobile-compatible wireless system. Due to the distinct layers of MISSA, it creates exclusive code bars for four distinct VOCs over three concentration levels. This grants precise gas identification and concentration prognoses with excellent accuracy of 99.77% and 98.3%, respectively. The combination of MISSA with artificial intelligence (AI) suggests its potential as a successful wearable device for long-term daily VOC monitoring and assessment for personal health monitoring scenarios.     

Methods for the Assessment of NET Formation: From Neutrophil Biology to Translational Research

Several studies have indicated that a neutrophil extracellular trap (NET) formation, apart from its role in host defense, can contribute to or drive pathogenesis in a wide range of inflammatory and thrombotic disorders. Therefore, NETs may serve as a therapeutic target or/and a diagnostic tool. Here, we compare the most commonly used techniques for the assessment of NET formation. Furthermore, we review recent data from the literature on the application of basic laboratory tools for detecting NET release and discuss the challenges and the advantages of these strategies in NET evaluation. Taken together, we provide some important insights into the qualitative and quantitative molecular analysis of NETs in translational medicine today.     

综掘工作面机载式临时护顶机的设计研究

综掘巷道采用锚杆支护时,临时支护是必不可少的一道工序。临时支护的可靠与否直接影响到锚杆支护作业的安全。针对目前临时支护现状时常发生伤亡事故,阐述了机载式临时护顶机的主要结构、工作原理及其主要技术参数,介绍该设备井下使用的情况,并与窜管式临时支护方式进行对比。     

基于熵权的沥青路面使用性能多目标综合评价

针对传统路面使用性能评价方法主观性强以及已有模型存在缺陷等问题,分析了高速公路沥青路面使用性能评价指标体系,通过主成分分析（principal component analysis,PCA）法对评价指标进行降维处理,形成彼此相互独立的主成分,并选取主成分样本训练支持向量机（support vector machine,SVM）,构建主成分分析法与支持向量机相结合的PCA-SVM模型以评价高速公路沥青路面使用性能.最后,选用罗宁高速2016年实测路面使用性能数据对提出的模型进行验证,并与《公路技术状况评定标准》的评价结果进行对比,结果表明：PCA-SVM模型评价结果与《公路技术状况评定标准》评价结果相近,修补率较高的5个路段评价结果比《公路技术状况评定标准》评价结果低一个等级,该研究成果可为公路技术状况评定的改进提供借鉴.

     

基于核主元分析法和支持向量机的人耳识别

对人耳识别中若干关键问题进行了研究.介绍了两种人耳图像归一化处理的方法,即基于外耳轮廓长轴的线标记法和基于外耳轮廓起始点的点标记法,并对这两种方法进行了对比.在分析现有人耳识别方法不足的基础上,提出利用核主元分析法提取人耳图像的代数特征,再利用支持向量机分类模型进行人耳识别.在带有角度、光照变化的北京科技大学人耳图像库上得到的识别率为98.7%,表明了该识别方法的有效性以及利用人耳图像进行身份识别的可行性.     

关于用户兴趣度的判定树算法的优化

判定树学习有很多算法,着重研究对引入用户兴趣度参数的ID3算法在面对多值属性时的快速分类的优化,提高计算速度,节约计算时间。     

油门踏板预变形注射模设计

介绍了汽车油门踏板的结构特点,通过模流分析计算出变形量,将3D数据按变形量进行反变形设计,达到注射成型后塑件的设计尺寸。     

一种基于多Agent的织机控制系统

介绍了有关多Agent（智能体）理论,对基于多Agent思想的织机控制系统的设计思路及具体实施方案进行了探讨,为通用性、智能性织机控制系统的开发提供了一种新的设计思路。