A methodology for identifying similar days in air traffic flow management initiative planning

This article describes a methodology for selecting days that are comparable in terms of the conditions faced during air traffic flow management initiative planning. This methodology includes the use of specific data sources, specific features of calendar days defined using these data sources, and the application of a specific form of classification and then cluster analysis. The application of this methodology will produce results that enable historical analysis of the use of initiatives and evaluation of the relative success of different courses of action. Several challenges are overcome here including the need to identify the appropriate machine learning algorithms to apply, to quantify the differences between calendar days, to select features describing days, to obtain appropriate raw data, and to evaluate results in a meaningful way. These challenges are overcome via a review of relevant literature, the identification and trial of several useful models and data sets, and careful application of methods. For example, the cluster analysis that ultimately selects sets of similar days uses a distance metric based on variable importance measures from a separate classification model of observed initiatives. The methodology defined here is applied to the New York area, although it could be applied by other researchers to other areas.     

Image segmentation based on competitive learning

lmage segment is a primat.‘y step in image analysis of unexploded ordnance (UXO) detection by ground penetrating radar (GPR) sensor which is accompanied with a lot of noises and other elements that affect the recognition of real target size. In this paper we bring forward a new theory, that is, we look the weight sets as target vector sets which is the new cues in semi-automatic segmentation to form the final image segmentation. The experiment results show that the measure size of target with our method is much smaller than the size with other methods and close to the real size of target.     

研究性学习在英语教学中的探索

研究性学习是一种新的学习方式,它鼓励学生在学习过程中运用所学知识和技能亲自去体验和实践,进行创造性、发现性的学习,从而提高综合运用所学知识解决实际问题的能力．文章提出在大学美语教学中,更要充分发挥研究性学习的特点,提高学生的英语综合运用能力．     

基于自动深度学习盾构掘进姿态预测与控制

提出了基于自动深度学习（AutoDL）算法和多目标优化算法的结合可实现数据驱动的姿态偏差控制指导,用于盾构掘进姿态的预测与控制,以解决现有盾构掘进姿态预测中所面临的执行难度高、成本高、效率低等问题,可用于自动精准地预测盾构掘进姿态随着工程进展的动态变化趋势,并针对盾构机施工状态执行多目标优化算法,快速自动搜寻最优策略,实时调整合适的盾构操作参数,减少对于现场操作人员经验和主观判断的依赖。以上海市天然气主干管网崇明岛-长兴岛-浦东新区五号沟LNG站管道工程隧道A线工程为例,展示该算法框架的优越性。研究结果有助于降低深度学习进入盾构智能控制领域的门槛,推动智能盾构发展。     

深度学习在激光熔覆领域的发展与应用

介绍激光熔覆的基本原理、应用领域和深度学习模型的发展状况,阐述深度学习在激光熔覆领域的应用。深度学习利用神经网络模型监控激光熔覆过程,在熔覆工艺参数优化和熔覆过程质量控制等方面具有较好优势,可有效提升激光熔覆过程的精度和稳定性。     

交叉口Agent间的多遇协调策略及其参数影响分析

为信号控制的城市道路交叉口定义一个Agent结构模型,在分析相邻交叉口交通流关联关系的基础上,利用记忆因子δ、学习概率α、交叉口交通流变化概率β<,i>等参数阐述了交叉口Agent间的多遇协调流程.交叉口 Agent多遇协调采用部分获利协作策略,其交互策略更多地考虑在获利少于对方时候如何以更加协作的态度进行协调.利用记...     

校内生产性实训基地建设的探索

高职校内生产性实训基地是以科学发展观为统领,遵循高等职业教育改革与发展规律,结合自身工作实际,满足实训教学及能力培养要求,在创新实训基地建设模式,拓展校际之间、校企之间合作,引进企业管理和长线产品上铺垫生产性实训平台,以此推动专业实训基地建设与改革。     

关于改进JAVA教学方法的研究

目前,很多高校都开设了与JAVA相关的教学课程,其教学效果的好坏,直接影响到学生的程序设计能力和对JAVA知识的学习兴趣。本文结合多年的JAVA教学经验,从三个方面进行了论述：把讲授JAVA开发工具的使用作为培养学生学习兴趣的切入点;将JAVA知识体系融贯于JAVA实例之中,把实例教学作为培养学生学习兴趣的手段;把培养学生继续学习JAVA知识的能力作为教学的最终目标。并将三者在教学中有机地结合在一起,探索适合JA-VA教学的方法,收到了良好的效果。     

浅谈如何做好一体化教学实施

本文就目前高职院校一体化教学在具体实施中存在的问题进行了分析,并结合高职院校的实际情况,对如何搞好一体化教学工作进行了探讨。     

对大学生正确掌握乒乓球正手攻球击球点的实证研究

<正>手攻球是乒乓球运动各项技术中必须要掌握的重要技术,正手攻球稳定性及准确性能力提高是决定比赛胜负的基础和根本。本文拟以两个班对照教学实验,对此问题进行实证研究。经研究发现,在来球上升初期和下降初期击球是保证攻球稳定性和成功率的关键技术。