考虑船舶速度偏差的集装箱班轮运输货运收益鲁棒优化

由于受到天气、海况及航路环境等不确定因素的影响,船舶实际航速常常与计划航速产生偏差,导致班轮运输的航次加油和货物装运策略也产生变化,进而对船舶燃油成本和运费收入产生影响。因此,考虑速度偏差的集装箱班轮运输收益优化研究具有重要的现实意义。为了保证速度偏差不确定性下的各航段船舶航行用油安全,以船舶到达下一挂靠港的船期时间窗为约束,推导得出速度偏差最大情况下的船舶航段最大燃油消耗量显性函数,结合各加油港燃油价格及折扣差异,以及集装箱货物各O-D流量及运费率差异,以班轮运输航次收益鲁棒优化为目标,构建了速度偏差下的集装箱班轮运输收益混合整数非线性规划模型,设计分段线性割线逼近算法进行模型求解。以中国远洋海运集团有限公司的MEX航线数据为实际算例进行了分析验证,结果显示：考虑速度偏差的班轮运输加油和货物装运全局鲁棒优化策略能够有效地提高班轮运输航次货运收益;随着船舶燃油消耗系数和各航段速度偏差极值的增加,班轮运输航次燃油消耗量也随之增加,而航次运费收入和货运收益都将随之减少。研究结论可为船公司制定速度偏差下的集装箱班轮运输决策提供有益的参考。     

基于模糊识别算法的财务指标分析系统的设计

本文利用数据挖掘技术,设计了一个具有财务分析和数据挖掘功能的财务指标分析系统,以便为企业决策层提供快速的财务分析支持。在指标分析模块中使用了模糊识别算法。该算法是利用训练样本集的模糊识别矩阵计算出模糊聚类中心矩阵,再利用模糊聚类中心矩阵反算出测试样本集的最优模糊识别矩阵,从而克服了传统的模糊聚类算法只能聚类的缺点。     

基于多分类器动态选择与成本敏感优化集成的电信客户流失预测研究

针对不同样本在特征空间中具有不同的区域特性和不同分类算法之间的预测互补性,在电信客户流失预测理论基础上,融合多分类器动态集成理论和成本敏感学习理论,建立了电信客户流失多分类器集成预测的利润函数,并提出了一类新的基于多分类器动态选择与成本敏感优化集成的电信客户流失预测模型.首先使用K均值聚类法聚类训练样本成多个分区;接着使用NaiveBayes算法、多层感知机算法和J48算法在各分区样本上构建客户流失预测子分类器;最后使用改进人工鱼群算法分别对各分区的子分类器进行成本敏感优化集成.实验结果表明,所提出的基于多分类器动态选择与成本敏感优化集成模型的分类性能不仅优于由训练集全体样本所构建的3个单模型,也优于基于改进人工鱼群算法优化集成这3个单模型而得到的集成模型.     

大放卫星的人——克雷格·麦考

移动通信业的先驱克雷格·麦考 已在探索一种全新的传播In-ternet信息的方法:使用上百颗卫星,即Teledesic计划。 Teledesic计划的核心是,让288颗小卫星在距地面700公里的高度上飞行,远远低于传统卫星的飞行高度。     

基于WEKA平台的文本聚类研究与实现

文本聚类是文本挖掘领域的一个重要研究分支.是聚类方法在文本处理领域的应用.本文首先对基于空间向量模型的文本聚类过程做了较深入的讨论和总结.另外,本文回顾了现有的文本聚类算法,以及常用的文本聚类效果评价指标.在研究了已有成果的基础上,本文利用20Newsgroup文本语料库,针对向量空间表示模型,在开源的数据挖掘平台WEKA上实现了文本预处理和k-means聚类算法,并根据实际聚类效果,就文本表示、特征选择、特征降维等方面提出优化方案.     

分类中的类重叠问题及其处理方法研究

类重叠问题是数据挖掘与机器学习领域的瓶颈问题之一.如果其中还存在类不均衡问题时,情况变得更加复杂.有鉴于此,本文在已有文献基础上归纳了三种类重叠学习算法及提出一种新的方法:分隔法,并首次将支持向量数据描述算法用于实际数据的重叠样本识别,对类重叠问题及其与类不均衡问题的相互影响进行了系统研究.在真实数据上采用五种分类器的实验结果表明:1)多数情况下“分隔法”是表现最佳的类重叠学习算法;2)分隔法通常对基于分界面而非规则的分类器更为有效;3)分隔法在类不均衡问题中表现很好,当基础分类器为支持向量机时尤为突出.最后针对支持向量机的实验结果给出了理论分析.     

基于密度的最佳聚类数确定方法

确定数据集的正确聚类数目是聚类分析中的一个基础性难题。常用的聚类数确定方法通常依赖特定的聚类算法,且在数据集存在子簇群的情况下效果欠佳。本文提出一种新的最佳聚类数确定的指标,该指标着重于分析簇的几何结构,从数据对象分布密度的角度来度量类内紧密度与类间分离度。该指标对噪声不敏感并且可以识别数据集中的子簇群,在实际数据和合成数据上的实验结果表明,新指标的性能优于广泛使用的其他指标。     

基于模糊聚类的高新技术企业信用风险评估方法研究

本文针对高新技术企业的特点,应用粗糙集理论和模糊聚类方法,集成了一种信用风险评估方法。首先构建了高新技术企业信用风险评估指标体系,再应用粗糙集约简方法生成约简的指标集,继而运用模糊聚类方法对企业信用风险进行评估。在使用模糊ISO-DATA聚类算法时,加入"引路样本",这种改进的算法提高了评估的客观性。     

移动商务中面向客户细分的KSP混合聚类算法

数据挖掘技术中的聚类算法是解决客户细分问题的重要算法之一。为解决传统聚类算法在客户细分问题中分类精度较低、收敛速度较慢的问题,着重对比分析传统聚类算法中K-m eans、自组织映射网络和粒子群3种算法的不足,提出融合3种算法优点的混合型聚类算法,该算法利用K-m eans和自组织映射网络对初始聚类中心进行优化,结合粒子群优化和K-m eans优化聚类迭代过程,并在迭代优化过程中设计避免算法因早熟而停滞的机制。针对移动电子商务环境下的餐饮业客户细分问题,建立移动餐饮业客户细分模型,并利用混合型聚类算法、K-m eans、层级自组织映射网络和基于粒子群的K-m eans等4种算法对实际案例进行对比分析。研究结果表明,混合型聚类算法的聚类精度分别比其他3种算法高,同时还具有最快的收敛性能,更适用于客户细分问题。     

基于项目聚类和评分预测的协同过滤推荐算法研究

协同过滤算法是目前个性化推荐系统中应用最成功的推荐算法之一。目前协同过滤构建的用户-项目矩阵,一般是按用户对所有项目的评分构建,却没有考虑项目之间的分类情况,导致寻找的邻居集合可能不是最近邻居集合。针对此问题,本文提出基于项目聚类和评分预测的协同过滤推荐算法,该算法首先按商品聚类,将大矩阵按聚类的商品来进行子矩阵的计算,在子矩阵里进行兴趣度的测量,最后将在所有区域相似用户的推荐项目合并,成为该用户的最后推荐结果。实验证明新算法能够提高协同过滤推荐系统的推荐质量。     

具有周期特征的末阶段飞行员资源紧张度评估方法研究

近年来,随着我国航空业的快速发展,飞行员不足成了很多航空公司发展的制约因素。由于飞行员的飞行时间有周、月、年度飞行小时的约束,航空公司在执行每月第四周的航班计划时,往往由于飞行员前三周飞行时间不平衡,造成利用剩余的飞行员资源完成第四周计划航班任务的紧张状况。本文针对此问题,提出一种评估每月第四周飞行员资源紧张程度的方法。该模型首先按照带队机长可用的飞行小时,将带队机长分为若干类,以各类带队机长的投入数量为投入变量,利用历史数据分析航班任务和飞行员投入之间的关系。利用各类飞行员投入结构比例的约束,计算第四周可用带队机长飞行小时可完成的最大航班任务量,通过与计划航班任务量的比较,评估完成该航班计划的紧张程度,并根据结果在需要预警时,给出不同级别的预警。从而可以使得安全生产关口前移,提高航空公司的安全生产水平,同时为飞行员的培训提供依据。     

Questioning the reliability of market segmentation techniques

Market segmentation techniques and analyses have received considerable attention in the academic literature. This paper compares and contrasts the results from conducting cluster analysis and factor analysis on two data sets, one which is genuine and the other which has been randomly generated. This analysis helps test claims about the validity and effectiveness of such techniques and looks at the reliability of the output. The findings suggest that researchers need to be careful to carry out valid preliminary work before presenting their results. Without such preliminary work it may be difficult to justify the segments which emerge.     

基于ASRank和MMAS的蚁群算法求解飞机指派问题

本文将航班串的飞机指派问题归结为车辆路径问题,考虑连续航班串之间衔接时间、衔接机场的约束、每架飞机的总飞行时间约束,建立了带有飞行时间约束的车辆路径问题的混合整数规划模型。构造了蚁群系统算法,引入基于排序的蚂蚁系统和最大最小蚂蚁系统算法的信息素更新策略。选取某航空公司7组初始航班串集合进行测试,并对算法中的重要参数进行了分析。实验结果表明,本文设计的模型和算法可以有效地减少连续航班串之间的总衔接时间,在可接受的计算时间内获得满意解。     

基于机场繁忙程度的航班延误波及分析

航班延误是全球航空业面临的一大难题。航班运行过程中，对于执行航班环任务的飞机，机场繁忙程度直接影响飞机过站时间长短，进而影响航班离港延误程度。文中构建到港延误对离港延误的波及贝叶斯网络预测模型时，加入机场繁忙程度这一因素，以机场飞机起降架次作为刻画该因素的指标，并通过贝叶斯网络结构学习得到机场繁忙程度的影响关系图。10次10折交叉验证的结果表明，与直接用到港延误预测离港延误的模型相比，加入机场繁忙程度因素的模型能够更准确地预估航班延误波及情况。     

Current and Future Particulate‐Matter‐Related Mortality Risks in the United States from Aviation Emissions During Landing and Takeoff

Demand for air travel is projected to increase in the upcoming years, with a corresponding influence on emissions, air quality, and public health. The trajectory of health impacts would be influenced by not just emissions growth, but also changes in nonaviation ambient concentrations that influence secondary fine particulate matter (PM_2.5) formation, population growth and aging, and potential shifts in PM_2.5 concentration‐response functions (CRFs). However, studies to date have not systematically evaluated the individual and joint contributions of these factors to health risk trajectories. In this study, we simulated emissions during landing and takeoff from aircraft at 99 airports across the United States for 2005 and for a 2025 flight activity projection scenario. We applied the Community Multiscale Air Quality (CMAQ) model with the Speciated Modeled Attainment Test (SMAT) to determine the contributions of these emissions to ambient concentrations, including scenarios with 2025 aircraft emissions and 2005 nonaviation air quality. We combined CMAQ outputs with PM_2.5 mortality CRFs and population projections, and evaluated the influence of changing emissions, nonaviation concentrations, and population factors. Given these scenarios, aviation‐related health impacts would increase by a factor of 6.1 from 2005 to 2025, with a factor of 2.1 attributable to emissions, a factor of 1.3 attributable to population factors, and a factor of 2.3 attributable to changing nonaviation concentrations which enhance secondary PM_2.5 formation. Our study emphasizes that the public health burden of aviation emissions would be significantly influenced by the joint effects of flight activity increases, nonaviation concentration changes, and population growth and aging.     

Minimization of passenger takeoff and landing risk in offshore helicopter transportation: Models,approaches and analysis

Offshore petroleum industry uses helicopters to transport the employees to and from installations. Takeoff and landing represent a substantial part of the flight risks for passengers. In this paper, we propose and analyze approaches to create a safe flight schedule to perform pickup of employees by several independent flights. Two scenarios are considered. Under the non-split scenario, exactly one visit is allowed to each installation. Under the split scenario, the pickup demand of an installation can be split between several flights. Interesting links between our problem and other problems of combinatorial optimization, e.g., parallel machine scheduling and bin-packing are established. We provide worst-case analysis of the performance of some of our algorithms and report the results of computational experiments conducted on randomly generated instances based on the real sets of installations in the oil fields on the Norwegian continental shelf. This paper is the first attempt to handle takeoff and landing risk in a flight schedule that consists of several flights and lays ground for the study on more advanced and practically relevant models.     

Segmenting the commercial producer marketplace for agricultural inputs

A cluster analysis procedure was used to develop a market segmentation of U.S. crop and livestock farms with annual sales in excess of $100,000. The segments were developed based on the importance of six factors that producers evaluate when selecting input suppliers. The results indicate that four distinct segments exist: Convenience buyers, Balance buyers, Price buyers, and Performance buyers. Differences in preferences across these segments have important implications for the marketing strategies of agricultural input suppliers.     

Dual Objective Segmentation to Improve Targetability: An Evolutionary Algorithm Approach**

Cluster‐based segmentation usually involves two sets of variables: (i) the needs‐based variables (referred to as the bases variables), which are used in developing the original segments to identify the value, and (ii) the classification or background variables, which are used to profile or target the customers. The managers’ goal is to utilize these two sets of variables in the most efficient manner. Pragmatic managerial interests recognize the underlying need to start shifting from methodologies that obtain highly precise value‐based segments but may be of limited practical use as they provide less targetable segments. Consequently, the imperative is to shift toward newer segmentation approaches that provide greater focus on targetable segments while maintaining homogeneity. This requires dual objective segmentation, which is a combinatorially difficult problem. Hence, we propose and examine a new evolutionary methodology based on genetic algorithms to address this problem. We show, based on a large‐scale Monte Carlo simulation and a case study, that the proposed approach consistently outperforms the existing methods for a wide variety of problem instances. We are able to obtain statistically significant and managerially important improvements in targetability with little diminution in the identifiability of value‐based segments. Moreover, the proposed methodology provides a set of good solutions, unlike existing methodologies that provide a single solution. We also show how these good solutions can be used to plot an efficient Pareto frontier. Finally, we present useful insights that would help managers in implementing the proposed solution approach effectively.     

Is the Swiss Labor Market Segmented? An Analysis Using Alternative Approaches

Abstract. In this paper, three common empirical methods encountered in the segmentation literature are used in order to establish whether or not the Swiss labor market is segmented: (i) a hierarchical cluster analysis; (ii) a switching model with unknown regime; and (iii) an analysis of low‐wage mobility with a bivariate probit model with endogenous selection. According to method (i), segmentation can hardly be observed. Method (ii) shows that the Swiss labor market is dualistic in nature. Method (iii) reveals that a certain degree of persistence exists in low‐wage jobs. Whether or not the Swiss labor market is segmented thus depends on the choice of method, i.e. on the definition and understanding of segments. In any case, none of the methods used in this study point to the existence of a large and well‐defined secondary segment.     

考虑区域限制的卡车搭载无人机车辆路径问题研究

无人机参与配送是解决末端物流难题的重要途径之一。卡车搭载无人机协同配送模式,克服了无人机载重量小、续航时间短的弊端,成为无人机参与末端物流配送的重要形式之一。在疫区、灾区进行应急配送时,经常遇到由于道路毁坏或封锁、区域污染,导致部分路段车辆或无人机无法通行的情况。在非应急配送中,也可能存在车辆限行和空域禁飞等区域限制措施。区域限制给卡车搭载无人机物流配送路径优化问题带来了很大挑战。论文构建了区域限制条件下卡车搭载无人机车辆路径问题的混合整数线性规划模型,提出了一种结合最短路算法和禁忌搜索算法的混合算法,基于标准算例库设计测试集并进行测试实验,实验结果表明混合算法具有较好的计算性能。