Behavioral Decision-Making of Key Stakeholders in Public-Private Partnerships:A Hybrid Method and Benefit Distribution Study

Public-private partnerships(PPPs)have been used by governments around the world to procure and construct infrastructural amenities.It relies on private sector expertise and funding to achieve this lofty objective.However,given the uncertainties of project management,transparency,accountability,and expropriation,this phenomenon has gained tremendous attention in recent years due to the important role it plays in curbing infrastructural deficits globally.Interestingly,the reasonable benefit distribution scheme in a PPP project is related to the behavior decisionmaking of the government and social capital,aswell as the performance of the project.In this paper,the government and social capital which are the key stakeholders of PPP projects were selected as the research objects.Based on the fuzzy expected value model and game theory,a hybrid method was adopted in this research taking into account the different risk preferences of both public entities and private parties under the fuzzy demand environment.To alleviate the problem of insufficient utilization of social capital in a PPP project,this paper seeks to grasp the relationship that exists between the benefit distribution of stakeholders,their behavioral decision-making,and project performance,given that they impact the performance of both public entities and private parties,as well as assist in maximizing the overall utility of the project.Furthermore,four game models were constructed in this study,while the expected value and opportunity-constrained programming model for optimal decision-making were derived using alternate perspectives of both centralized decision-making and decentralized decision-making.Afterward,the optimal behavioral decision-making of public entities and private parties in four scenarios was discussed and thereafter compared,which led to an ensuing discussion on the benefit distribution system under centralized decision-making.Lastly,based on an example case,the influence of different confidence levels,price,and fuzzy uncertainties of PPP projects on the equilibrium strategy results of both parties were discussed,giving credence to the effectiveness of the hybrid method.The results indicate that adjusting different confidence levels yields different equilibriumpoints,and therefore signposts that social capital has a fair perception of opportunities,as well as identifies reciprocal preferences.Nevertheless,we find that an increase in the cost coefficient of the government and social capital does not inhibit the effort of both parties.Our results also indicate that a reasonable benefit distribution of PPP projects can assist them in realizing optimum Pareto improvements over time.The results provide us with very useful strategies and recommendations to improve the overall performance of PPP projects in China.

Data complexity-based batch sanitization method against poison in distributed learning

The security of Federated Learning(FL)/Distributed Machine Learning(DML)is gravely threatened by data poisoning attacks,which destroy the usability of the model by contaminating training samples,so such attacks are called causative availability indiscriminate attacks.Facing the problem that existing data sanitization methods are hard to apply to real-time applications due to their tedious process and heavy computations,we propose a new supervised batch detection method for poison,which can fleetly sanitize the training dataset before the local model training.We design a training dataset generation method that helps to enhance accuracy and uses data complexity features to train a detection model,which will be used in an efficient batch hierarchical detection process.Our model stockpiles knowledge about poison,which can be expanded by retraining to adapt to new attacks.Being neither attack-specific nor scenario-specific,our method is applicable to FL/DML or other online or offline scenarios.

Conjunctive Use of Engineering and Optimization in Water Distribution System Design

Water Distribution Systems (WDSs) design and operation are usually done on a case-by-case basis. Numerous models have been proposed in the literature to solve specific problems in this field. The implementation of these models to any real-world WDS optimization problem is left to the discretion of designers who lack the necessary tools that will guide them in the decision-making process for a given WDS design project. Practitioners are not always very familiar with optimization applied to water network design. This results in a quasi-exclusive use of engineering judgment when dealing with this issue. In order to support a decision process in this field, the present article suggests a step-by-step approach to solve the multi-objective design problem by using both engineering and optimization. A genetic algorithm is proposed as the optimization tool and the targeted objectives are: 1) to minimize the total cost (capital and operation), 2) to minimize the residence time of the water within the system and 3) to maximize a network reliability metric. The results of the case study show that preliminary analysis can significantly reduce decision variables and computational burden. Therefore, the approach will help network design practitioners to reduce optimization problems to a more manageable size.

基于批数据过采样的中医临床记录四诊描述抽取方法

中医临床记录四诊描述抽取对中医临床辨证论治的提质增效具有重要的应用价值,然而该任务尚有待探索,类别分布不均衡是该任务面临的关键挑战之一。该文围绕中医临床记录四诊描述抽取任务展开研究,首先构建了中医临床四诊描述抽取语料库;然后基于无标注的中医临床记录微调通用预训练语言模型实现该模型的领域适应;最后利用小规模标注数据,采用批数据过采样算法,完成中医临床记录四诊描述抽取模型的训练。实验结果表明,该文提出的抽取方法的总体性能均优于对比方法,并且与对比方法的最优结果相比,该文方法将少见类别的抽取性能F1值平均提升了2.13%。

带有动态到达工件的分布式柔性作业车间调度问题研究

分布式柔性作业车间调度是生产调度的1个重要分支,工件的动态到达作为实际生产中的1种常见扰动情况,进一步增加了作业车间调度问题的复杂性和不确定性。针对带有工件动态到达的分布式柔性作业车间调度问题(DA-DFJSP),提出1种分批调度策略,将原本的动态调度问题转化成一系列连续调度区间上的静态调度问题,构建以最大完工时间为优化目标的混合整数规划模型;在此基础上,结合问题特征采用批次、工厂、工序、机器的4层染色体编码及快速贪婪搜索插入的解码方式改进遗传算法,同时引入多种交叉、变异算子来增强染色体的多样性;最后,基于FJSP标准算例构建DA-DFJSP测试算例进行仿真对比实验,验证所提策略和改进算法的求解优势。结果表明:相较于传统的重调度策略和改进前的遗传算法,采用分批调度策略和改进的遗传算法(IGA)所求调度方案具有更短的完工周期、更均匀的工厂加工负荷及更高的设备工作效率,IGA与分批调度策略之间有高度的契合性,能够有效提升生产效率。

联合双维度用户调度的自适应联邦边缘学习

联邦边缘学习无需传输本地数据,在保护用户隐私的基础上,大大降低了上行链路压力。联邦边缘学习通过智能边缘设备,利用本地数据集训练局部模型后上传模型参数至中心服务器;中心服务器聚合本地上传的局部模型参数形成全局模型后进行更新,然后将更新后的模型下发给智能边缘设备开始新一轮迭代。但是局部模型精度以及局部模型训练时间,对全局模型聚合以及模型更新过程会产生重大影响。因此提出自适应动态批量梯度下降策略,在局部模型训练过程中自动调整梯度下降抽取的批量大小,优化联邦学习的局部模型精度及收敛速度。针对用户数据的非独立同分布特性,设计一种联合双维度用户调度策略的自适应动态批量梯度下降算法,通过收敛时间和数据多样性进行双维度约束。经MNIST数据集、fashion MNIST数据集和CIFAR-10数据集的训练测试,算法在有效降低聚合等待时间的同时能够进一步提高全局模型精度和收敛速度。与固定批量为64、128、256的梯度下降算法相比,该算法的全局模型精度在运行100 s时提升分别为32.4%、45.2%、87.5%。

A Pioneering Experimental Study on the Batch Crystallization of the Citric Acid Monohydrate

Citric acid is an important organic substance whose marketing concerns various fields. Nevertheless, until 1997 the scientific literature reported little information about the process of crystallization by cooling through which the commercial product is obtained. In particular, the available studies were aimed to investigate only the kinetics of nucleation and crystal growth neglecting some effective aspects of the industrial crystallization in mechanically stirred apparatus. In order to fill that sci-tech gap, the Department of Chemical Engineering at the University ＂La Sapienza＂ of Rome decided to lead a long and meticulous experimental research on the crystallization in discontinuous （batch） of CAM （citric acid monohydrate） in the allotropic form that is stable at room temperature. Due to the number of people involved in that pioneering work, carried out in the historic laboratories of＂La Sapienza＂ （Faculty of Engineering）, and motivated by the publication of related M.Sc. dissertations and research papers, such collective effort was called ＂School of Industrial Crystallization＂. Among the graduate students in Chemical Engineering that 17 years ago participated in that fruitful experience there was also the author who, under the supervision of Prof. Barbara Mazzarotta, had the specific task of assessing the effects on CAM of changing the crystallization operating conditions until their optimization; the achievements are briefly illustrated in this paper.

基于Hough变换的企业财务重复数据批量剔除方法

企业财务数据在空间分布上具有一定的不规则性,导致对重复数据剔除的完整性较低,为此,提出基于Hough变换的企业财务重复数据批量剔除方法。首先,采用Hough变换对企业财务数据进行预处理,将累加器单元所有直线形成的峰值作为数据理想的期望值,根据期望变换结果区分在Hough变换下数据的位置变化幅值,沿原始数据空间分布实现对数据的划分。考虑到企业财务数据对应空间直线的边缘即为直线形成的累加器单元形状,因此对沿着正弦曲线法线方向的累加值进行增强处理,得到放大后的峰值信息。之后,采用ASCA实现对重复数据的筛选,用二分极值法将距离Hough变换后峰值最远的位置作为聚类中心,将DBI指数完全一致的数据聚类,保留聚类中的唯一数据,其余进行批量删除,完成操作后计算新的聚类中心,重复计算直至峰值累加器单元的财务数据处于均匀分布状态。测试结果表明,设计方法空间缩减率可达到91.0%,删除数据的准确性可达到96.12%,实现了对重复数据的较完整剔除。

服务时间累积量控制的Geo^(X)/G/1排队的稳态队长

考虑服务台的启动由服务台闲期所有到达的服务时间累积之和(即顾客即将被服务的时间累积量)控制的离散时间批到达Geo^(X)/G/1排队服务系统。顾客成批到达系统,当顾客的被服务时间累积量超过某个非负整数D时,服务台立刻为顾客提供服务直到忙期结束(此策略被称为D策略)。该模型可为无线传感网络的从业者提供一些理论基础。首先,在准备工作中,讨论了忙期开始时刻的顾客数、服务时间累积量以及服务台的忙期和闲期分布;然后,通过闲、忙期到达顾客的分类和概率分析方法,研究了系统的离去时刻稳态队长和任意时刻n^(+)的稳态队长分布的概率母函数。作为特例,分析得到了离散时间Geo^(X)/G/1排队和D策略离散时间Geo/G/1排队的队长分布结果。最后,模拟分析了一类无线传感节点,并在数值上获得了节点的最低能耗。

基于机器学习的配电网监控信息批处理方法

智能配电网系统存在计算机延迟、信息批处理存储负荷较低的问题,为此提出基于机器学习的配电网监控信息批处理方法。利用HDFS分布式文件系统衡量数据传输记录,搭建配电网监控信息调度集群;测算计算机系统的延迟性,利用数据响应任务填补系统延迟时间;规划监控信息接口的处理模式,对机器学习的大规模流计算进行区域划分,构建监控信息批处理模型;对信息调度集群进行特征化划分,增加并统计信息批处理堆积节点;利用配电网监控信息堆积特征扩展规律,通过云计算服务层次制定智能分析机器的学习路径;智能分析机器训练分析配电监控信息系统样本,采集配电网监控信息堆积特征;利用机器学习标记有效监控信息和有效事件信息,计算待聚集处理的遥测信息数量;通过HDFS规划遥测协调节点的布局,对监控信息的存储空间进行划分;利用遥控元数据库对存储空间进行监控,利用管理器完成堆积V002配电网监控信息的批处理与存储。实验结果表明,基于机器学习的配电网监控信息批处理方法在信息集群处理中的存储负荷率最高为95%,其配电网监控信息批处理效果得到保证。

空港物流暂存区货物分拨配送问题研究

空港物流地面暂存区爆仓和配送车辆装载率低的问题一直亟待解决。基于航空公司暂存区货物入库和出卡流程,考虑航班的到达时间、货物目的地、货物件数和货物重量等信息,以货物在暂存区的滞留时间成本和配送车辆运输成本之和最小化为目标建立混合整数规划模型,并设计了聚类算法。基于MU航空公司的真实数据生成算例进行模拟计算,并与该航空公司实际分拨配送方案进行对比,验证了所提方法的有效性。结果显示,优化后货物分拨配送方案的货物在暂存区的平均滞留时间减少了32%,配送车辆装载率提升了32%,总成本降低了31%,为解决空港物流地面暂存区爆仓问题提供了决策支持。

基于笛卡尔乘积图上Sobolev平滑的时变图信号分布式批量重构

针对大规模网络数据的重构问题,该文以图信号处理(GSP)理论为基础,提出一种基于笛卡尔乘积图上Sobolev平滑的分布式批量重构算法(DBR-SSC)。该算法首先按时间顺序将时变图信号划分为多个信号段,并利用笛卡尔积将每一段内各时刻的图建模为乘积图;然后利用笛卡尔乘积图上的Sobolev差分平滑,将每一段的信号重构问题归结为优化问题;最后设计具有高收敛速度的分布式算法求解该优化问题。采用两种现实世界的数据集进行仿真实验,实验结果表明所提算法重构误差低并具有高收敛速度。

基于仿真的多品种小批量生产资源负荷预测

生产计划的制订需充分考虑生产资源的负荷情况。在多品种小批量生产模式下,工序加工时间的不确定性以及生产加工质量等因素的影响会导致加工过程中实际资源负荷与计划之间存在偏差。针对上述问题,对实际生产所需的资源进行负荷预测。考虑工序历时、加工不合格率以及设备使用前准备时间对资源负荷的影响,提出一种资源负荷预测模型并应用蒙特卡洛仿真方法,以概率分布的形式描述工序加工历时的不确定性。通过仿真验证了预测模型的可行性,实现了对生产资源负荷较为符合实际的预测,为制定生产计划提供有效参考。

基于多项分布的弹药多批次试验命中概率估计

弹药命中概率试验通常分多个批次进行,某一批次试验结束后会对弹药物理性能进行改进,各批次试验命中概率呈现动态变化特点。每批次试验前会获取到该批次弹药命中概率仿真实验信息,充分利用各批次射击试验信息与仿真实验信息是弹药命中概率估计的重点。考虑到弹药命中目标不同区域造成的毁伤程度不同,对原有二项分布命中概率试验进行改进,改用多项分布进行描述。引入体系贡献度衡量各批次试验对弹药命中概率估计的影响程度,并结合贝叶斯方法对各批次试验信息进行融合,通过构建顺序约束关系鉴定弹药命中性能改进效果,并提出基于多项分布的弹药多批次试验命中概率估计方法。研究结果表明,该方法相比二项分布方法能够细致描述弹药命中性能,并科学融合各批次射击试验信息与仿真实验信息,为弹药命中概率估计提供借鉴。

ArcPy脚本在新型基础测绘中的应用研究

在日常数据处理过程中,由于空间数据的数据量较大,在对其进行重复性数据处理工作中,工作人员需进行大量手动操作且容易出错。空间数据的自动批处理可以大幅度降低数据处理人员的工作强度,提高工作效率。结合山东省新型基础测绘生产项目,使用ArcPy脚本对一些重复性的工作进行批处理,实现数据分发、影像挑选、数据合并等的批量操作,为ArcPy脚本在空间数据自动批处理中的应用提供一种可行且高效的方案。

改进漏斗式喷气静电纺丝装置及其批量制备有序纳米纤维机理研究

提出了一种改进的漏斗式喷气静电纺丝装置,其由一个钛合金锥形喷嘴、一个上壁面具有多个凹槽的钛合金储液池和一个用作收集器的铜丝滚筒组成,可以获得更高产量的有序纳米纤维。为了研究改进的漏斗式喷气静电纺丝装置批量制备纳米纤维的机理,分别模拟了其与原有的漏斗式喷气静电纺丝装置的电场分布,并通过理论和实验分析了储液池结构和纺丝电压对纳米纤维形貌和产量的影响,实验数据与模拟结果相一致,表明改进的漏斗式喷气静电纺丝装置纺丝性能更好,可获得更高质量和产量的有序纳米纤维,且45 kV为最优纺丝电压。

一起配电变压器批次性故障的诊断与分析

本文中作者通过绝缘电阻试验、绕组直流电阻试验,并结合吊芯检查,分析了一起因绕组焊接质量缺陷导致的配电变压器批次性故障,并针对故障原因提出了防范措施。

基于云平台的批量话单快速解码方法

针对当前运营商业务支撑系统的处理话单能力,特别是对用户批量漫游话单解码处理慢的现状,探索了利用分布式计算框架搭建的云计算能力解码的方法。在服务器内存中建立分层处理逻辑,将解码需要的配置信息、索引信息等关键数据放入内存中,同时建立多节点部署的集群处理架构,采用云计算方式对批量话单进行并行解码处理,实时输出处理结果。实验表明该方法可有效提升批量话单解码的效率,有效提升运营商对外服务的能力。

^(239)Pu在泥岩和砂岩上的吸附分配系数

通过静态批式法研究放射性核素中具有代表性的核素239Pu在泥岩和砂岩中的吸附情况以及多种影响Pu在泥岩、砂岩中吸附性能的因素。实验测定了泥岩、砂岩中239Pu的分配系数,考察了岩石粒径、水相pH值、固液比和溶液离子对吸附分配系数的影响。结果表明,Pu在泥岩和砂岩中的分配系数分别为120 mL/g和1.60×104mL/g,砂岩对Pu有较强的吸附性;239Pu的吸附分配系数随岩石粒径的减小、水相pH的增大、固液比的增大而增大;硝酸溶液体系中,239Pu在岩石中的吸附性强于模拟地下水体系。

序批式反应器的好氧颗粒污泥特性研究

对序批式反应器中好氧颗粒活性污泥的形成过程、处理性能和颗粒分布特性进行了研究。结果表明,不同操作条件下产生了结构形态不同的颗粒污泥,沉降时间是颗粒污泥形成的主要因素,有机负荷对颗粒污泥的结构有一定影响。颗粒污泥反应器对溶解性COD的去除率可达90%,对氨氮的去除率为24%。颗粒污泥在反应器中分布不均匀,反应器底部MLSS高达8 g/L,SVI为34 mL/g,随着反应器高度的增加则MLSS值降低、SVI值提高;此外较大的颗粒污泥聚集在反应器的底部,较小的颗粒污泥和絮体分布在反应器的中上部。