Machine Learning-Based Two-Stage Data Selection Scheme for Long-Term Influenza Forecasting

One popular strategy to reduce the enormous number of illnesses and deaths from a seasonal influenza pandemic is to obtain the influenza vaccine on time. Usually, vaccine production preparation must be done at least six months in advance, and accurate long-term influenza forecasting is essential for this. Although diverse machine learning models have been proposed for influenza forecasting, they focus on short-term forecasting, and their performance is too dependent on input variables. For a country’s long-term influenza forecasting, typical surveillance data are known to be more effective than diverse external data on the Internet. We propose a two-stage data selection scheme for worldwide surveillance data to construct a long-term forecasting model for influenza in the target country. In the first stage, using a simple forecasting model based on the country’s surveillance data, we measured the change in performance by adding surveillance data from other countries, shifted by up to 52 weeks. In the second stage, for each set of surveillance data sorted by accuracy, we incrementally added data as input if the data have a positive effect on the performance of the forecasting model in the first stage. Using the selected surveillance data, we trained a new long-term forecasting model for influenza and perform influenza forecasting for the target country. We conducted extensive experiments using six machine learning models for the three target countries to verify the effectiveness of the proposed method. We report some of the results.     

Interval analysis of mode shapes to identify damage in beam structures

Various damage detection methods have been proposed by several researchers in the past few decades. Amongst them, the efficiency of mode shapes in detecting damage has been demonstrated by many researchers when further processed. In most cases, the processing involves expansion or reduction of the mode shape data. However, vital information that are damage-prints are often lost during processing of the mode shape data. In addition, most of these processes involve long and complex computation, thus, leading to inaccurate damage identification. In this study, a simple and fast damage identification technique is proposed to identify damage in beam structures. Interval analysis is applied to the mode shapes of a beam structure in the damaged and undamaged states. The interval situations of each of the beam's segment via mode shape are derived to obtain the upper and lower bounds and the derived bounds are compared. To establish a relationship for identify the damaged point, a possibility of damage existence is defined for each segment of the beam structure. The mode shape increment is defined as the increase in the mode shape value. Furthermore, a damage measure index that provide enhance damage information is obtained as the product of the possibility of damage existence and mode shape increment. A numerical model of a simply supported steel beam is applied to demonstrate this method by imposing damage through thickness reduction of elements in segments. In addition, a parametric analysis is carried out to evaluate noise effect by considering varying damage severities and different noise levels. The results showed that this method is simple and provides considerable accurate results.     

Does the method for Military Utility Assessment of Future Technologies provide utility?

The Military Utility Assessment of Future Technologies (MUAFT) method was developed as a cost-efficient alternative to methods such as NATO's Disruptive Technology Assessment Games, to be used as a part of the Swedish Armed Forces' long-term capability development process. The question addressed in this study is whether MUAFT can be considered to have validity in its context and thus if it has potential to be useful to other small to medium size states. The analysis was based on an operationalization of Clark's framework for science and technology intelligence analysis, combined with a military capability centric view of military utility. MUAFT reports from 2012 to 2018 were reviewed in terms of how they satisfy five key criteria. The study shows that MUAFT provides utility, if used by a suitably composed group of experts, who are aware of the method's limitations. The limitations mainly originate from a lack of explicit support for assessing the impact of forces for change, other than technological forces, on military capability development. The expert group serves as the synthesizing bridge between technology forecasts and military utility assessments. Therefore, comprehensive expertise is needed in various military technology specialisations, in the sponsor's military capabilities and in subjects necessary to master in order to assess other influential societal forces for change.     

基于时间延迟的多类型维修与经济生产批量联合优化研究

针对目前未能较好的综合考虑多维修类型与经济生产批量联合优化的问题，首先考虑多类型维修关系，基于时间延迟理论求出故障和缺陷次数的表达式；其次，在此基础上，综合考虑生产费用和维修费用的基础上，构建了多类型维修和经济生产批量联合优化模型，以单位时间内总费用最小为优化目标，获得最优检查间隔期和经济生产批量；最后，通过算例分析验证了模型的有效性，说明了第一类缺陷检查次数的多少对费用和经济生产批量影响不大。     

Extended vertical lists for temporal pattern mining from multivariate time series

In this paper, the problem of mining complex temporal patterns in the context of multivariate time series is considered. A new method called the Fast Temporal Pattern Mining with Extended Vertical Lists is introduced. The method is based on an extension of the level‐wise property, which requires a more complex pattern to start at positions within a record where all of the subpatterns of the pattern start. The approach is built around a novel data structure called the Extended Vertical List that tracks positions of the first state of the pattern inside records and links them to appropriate positions of a specific subpattern of the pattern called the prefix. Extensive computational results indicate that the new method performs significantly faster than the previous version of the algorithm for Temporal Pattern Mining; however, the increase in speed comes at the expense of increased memory usage.     

基于翻领松量结构模型的翻折领结构设计方法

为解决翻领松量计算精准性及翻折领通用结构设计方法的问题,提出了一种基于翻领松量结构模型的翻折领结构设计方法。通过构建翻领松量结构模型确定了翻领松量计算系数,对6种不同状态下的翻领间隙量变化进行数据统计分析,依据翻领间隙回归系数模型确定了翻领松量计算给定常数,运用SPSS统计软件对实验数据进行了相关性分析、回归分析、配对样本T检验,并推导出翻领松量计算公式。实验结果表明:使用该公式计算的翻领松量提高了翻折领结构造型的准确性,基于翻领松量结构模型的翻折领结构设计方法可以满足不同翻折领领型结构设计要求。     

Deep reinforcement learning for selecting demand forecast models to empower Industry 3.5 and an empirical study for a semiconductor component distributor

A semiconductor distributor that plays a third-party role in the supply chain will buy diverse components from different suppliers, warehouse and resell them to a number of electronics manufacturers with vendor-managed inventories, while suffering both risks of oversupply and shortage due to demand uncertainty. However, demand fluctuation and supply chain complexity are increasing due to shortening product life cycle in the consumer electronics era and long lead time for capacity expansion for high-tech manufacturing. Focusing realistic needs of a leading distributor for semiconductor components and modules, this study aims to construct a UNISON framework based on deep reinforcement learning (RL) for dynamically selecting the optimal demand forecast model for each of the products with the corresponding demand patterns to empower smart production for Industry 3.5. Deep RL that integrates deep learning architecture and RL algorithm can learn successful policies from the dynamic and complex real world. The reward function mechanism of deep RL can reduce negative impact of demand uncertainty. An empirical study was conducted for validation showing practical viability of the proposed approach. Indeed, the developed solution has been in real settings.     

Wire-free electricity: Insights from a techno-futuristic exploration

Wireless Power Transmission is an emerging technology that enables the transmission of electricity without the use of artificial conducting mediums. Lack of significant technological breakthroughs widely affected its progress. In this paper, we analyze the early progress of WPT and attempt to forecast its growth using Pearl curve model. Fisher-Pry substitution model analysis indicates that technology substitution of wired transmission has just begun and half the substitution will be over by 2028. We conducted patent landscaping to identify hot domains and specific technology areas. Implications for technology developers and manufacturers, electric utility providers and national policymakers are also identified.     

On auto- and cross-interdependence in interval field finite element analysis

This paper discusses the concepts of auto- and cross-interdependence in interval field finite element analysis. In classic interval analysis, independent intervals are used to construct hyper-rectangular input spaces that correspond to the bounded uncertainty that is present on some model parameters. This is a direct result from the inability of modeling interdependence. Such assumption of complete independence might prove in some cases to be highly over-conservative. A first example is the modeling of spatial uncertainty, where the interdependence is governed by allowable spatial gradients of field realizations. Secondly, interdependence can also occur in case uncertainty in several structural quantities has the same root cause (eg, the manufacturing process). Recent work by the authors introduced concepts for modeling dependence between intervals in a spatial and multivariate context. However, it is unclear how an analyst has to deal with multiple quantities that have a spatial uncertainty component and are furthermore interdependent. This paper presents an approach to link multiple interval fields using recently introduced convex hull pair constructions and inverse distance weighting interpolation. Two case studies to illustrate the new methodology are included, proving the flexibility of the methodology in the modeling of auto- and cross-interdependence between multiple interval scalars and/or interval fields.     

Improving basic relationships of pipe hydraulics

This communication presents the approaches set up for processing spinner flowmeter well logs in vertical wells with a single fluid phase, which is the most widely used in assessing wells productivity. These focus on improving the pipe hydraulics relationships so that the different fluid inputs throughout the well can be quantified. Since vertical flow inside wells varies with depth between laminar flows (very low Reynolds number, i.e. Re < 10³) and turbulent (Re > 4·10³) the aim has been to reduce the uncertainty in the transition interval. Starting from bibliographical data and/or well-known formulas for laminar and for turbulent flow, several continuous relationships have been developed for any regime: 1) an expression for the radial distribution of velocity inside the pipeline (velocity profile) was developed. 2) A relationship between the average velocity and the velocity at the axis (velocity factor) was created. 3) A third equation was generated to obtain the friction factor in smooth pipes (and starting from this, a new explicit equation for rough pipes). The purpose has been to have a set of empirical expressions of easy and continuous application for any regime, as an alternative to the use of computer simulations.