Application of an EWMA combining technique to the prediction of currency exchange rates

Financial forecasting is an important and challenging task for both academic researchers and business practitioners. The recent trend to improve the prediction accuracy is to combine individual forecasts using a simple average or weighted average where the weight reflects the inverse of the prediction error. In the existing combining methods, however, the errors between actual and predicted values are equally reflected in the weights regardless of the time order in a forecasting horizon. In this paper, we propose a new approach where the forecasting results of Generalized AutoRegressive Conditional Heteroskedastic (GARCH), neural network, and random walk models are combined based on a weight that reflects the inverse of the exponentially weighted moving average of the Mean Absolute Percentage Error (MAPE) of each individual prediction model. The results of an empirical study indicate that the proposed method has a better accuracy than the GARCH, neural network, and random walk models, and also combining methods based on using the MAPE for the weight.     

The use of time series forecasting in zone order picking systems to predict order pickers’ workload

In order to differentiate from competitors in terms of customer service, warehouses accept late orders while providing delivery in a quick and timely way. This trend leads to a reduced time to pick an order. This paper introduces workload forecasting in a warehouse context, in particular a zone picking warehouse. Improved workforce planning can contribute to an effective and efficient order picking process. Most order picking publications treat demand as known in advance. As warehouses accept late orders, the assumption of a constant given demand is questioned in this paper. The objective of this study is to present time series forecasting models that perform well in a zone picking warehouse. A real-life case study demonstrates the value of applying time series forecasting models to forecast the daily number of order lines. The forecast of order lines, along with order pickers’ productivity, can be used by warehouse supervisors to determine the daily required number of order pickers, as well as the allocation of order pickers across warehouse zones. Time series are applied on an aggregated level, as well as on a disaggregated zone level. Both bottom-up and top-down approaches are evaluated in order to find the best-performing forecasting method.     

Multiscale modeling of composite materials: a roadmap towards virtual testing

A bottom-up, multiscale modeling approach is presented to carry out high-fidelity virtual mechanical tests of composite materials and structures. The strategy begins with the in situ measurement of the matrix and interface mechanical properties at the nanometer-micrometer range to build up a ladder of the numerical simulations, which take into account the relevant deformation and failure mechanisms at different length scales relevant to individual plies, laminates and components. The main features of each simulation step and the information transferred between length scales are described in detail as well as the current limitations and the areas for further development. Finally, the roadmap for the extension of the current strategy to include functional properties and processing into the simulation scheme is delineated.     

Forecasting container throughput using aggregate or terminal-specific data? The case of Tanjung Priok Port,Indonesia

We propose a new approach to forecasting total port container throughput: to generate forecasts based on each of the port’s terminals and aggregate them into the total throughput forecast. We forecast the demand for total container throughput at the Indonesia’s largest seaport Tanjung Priok Port, employing SARIMA, the additive and multiplicative Seasonal Holt-Winters (MSHW) and the Vector Error Correction Model (VECM) on the monthly port and individual terminal container throughput time series between 2003 and 2013. The performance of forecasting models is evaluated based on mean absolute error and root mean squared error. Our results show that the MSHW model produces the most accurate forecasts of total container throughput, whereas SARIMA generates the worst in-sample model fit. The VECM provides the best model fits and forecasts for individual terminals. Our results report that the total container throughput forecasts based on modelling the total throughput time series are consistently better than those obtained by combining those forecasts generated by terminal-specific models. The forecasts of total throughput until the end of 2018 provide an essential insight into the strategic decision-making on the expansion of port’s capacity and construction of new container terminals at Tanjung Priok Port.     

Research in Production Planning and Inventory Control: Current Trends and Future Directions

During the past three decades important progress has been made in research on the planning and control of production and distribution systems. A conceptual framework has gradually developed that provides an overall structure for integrating individual planning and control decisions. Key elements in this structure include: forecasting the demand for individual products/services and product/ service categories; planning aggregate production, inventory, and capacity levels; scheduling the production of individual products, components, and raw materials; and controlling the inventory levels of individual product items. Furthermore, substantial progress had been made in the development of theory regarding decision making in each of these areas. For example, inventory theory provides an understanding of the basic trade-offs in the control of inventories. Likewise, scheduling theory provides a fundamental understanding of the performance impact of sequencing decisions. Finally, important progress has been made in the development and the installation of computer-based systems for planning and controlling production/distribution operations using the methods developed in theoretical research.     

Solving the grocery backroom sizing problem

Backrooms are an important echelon of the retail supply chain. However, research focus has been mostly targeted to optimise both distribution centres and stores' sales area. In this paper, we propose two mathematical programming formulations to solve the grocery backroom sizing problem. This problem consists of determining the dimension of each storage department in the backroom area to optimise its overall efficiency. The first formulation is a bottom-up approach that aims to reduce the backroom life-cycle costs by determining the optimum floor space and storage height for each department. The second is a top-down approach based on Data Envelopment Analysis (DEA), which determines the efficient level of storage floor space for each backroom department, based on a comparison with the benchmarks observed among existing stores. Each approach has distinct characteristics that turn the models suitable for different retail contexts. We also describe the application of the proposed approaches to a case study of a European retailer. The application of this methodology in the design process demonstrated substantial potential for space savings (6% for the bottom-up model and 16% for the top-down model). This space reduction should either allow higher revenues in the sales area and/or lower backroom-related costs.     

Bottom-up assembly of large-area nanowire resonator arrays

Directed-assembly of nanowire-based devices will enable the development of integrated circuits with new functions that extend well beyond mainstream digital logic. For example, nanoelectromechanical resonators are very attractive for chip-based sensor arrays because of their potential for ultrasensitive mass detection. In this letter, we introduce a new bottom-up assembly method to fabricate large-area nanoelectromechanical arrays each having over 2,000 single-nanowire resonators. The nanowires are synthesized and chemically functionalized before they are integrated onto a silicon chip at predetermined locations. Peptide nucleic acid probe molecules attached to the nanowires before assembly maintain their binding selectivity and recognize complementary oligonucleotide targets once the resonator array is assembled. The two types of cantilevered resonators we integrated here using silicon and rhodium nanowires had Q-factors of approximately 4,500 and approximately 1,150, respectively, in vacuum. Taken together, these results show that bottom-up nanowire assembly can offer a practical alternative to top-down fabrication for sensitive chip-based detection.     

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.     

Photopatterning of actin filament structures

We report a general strategy for spatiotemporal control of actin polymerization in vitro using photoactivatable actin. Caged actin was synthesized by chemically modified lysine residues on monomeric actin and released with focused ultraviolet (UV) illumination. Epifluorescence microscopy revealed nucleation and elongation of individual actin filaments (8 nm in diameter) after localized release of caged actin. We also used this strategy to generate branched filament structures by releasing caged actin in the presence of actin binding proteins. Controlled self-assembly of actin filaments represents a versatile "bottom-up" technique for constructing structural building blocks and functional templates for nanoscience applications.     

Micro- and Nanopatterning of Halide Perovskites Where Crystal Engineering for Emerging Photoelectronics Meets Integrated Device Array Technology

Tremendous efforts have been devoted to developing thin film halide perovskites (HPs) for use in high-performance photoelectronic devices, including solar cells, displays, and photodetectors. Furthermore, structured HPs with periodic micro- or nanopatterns have recently attracted significant interest due to their potential to not only improve the efficiency of an individual device via the controlled arrangement of HP crystals into a confined geometry, but also to technologically pixelate the device into arrays suitable for future commercialization. However, micro- or nanopatterning of HPs is not usually compatible with conventional photolithography, which is detrimental to ionic HPs and requires special techniques. Herein, a comprehensive overview of the state-of-the-art technologies used to develop micro- and nanometer-scale HP patterns, with an emphasis on their controlled microstructures based on top-down and bottom-up approaches, and their potential for future applications, is provided. Top-down approaches include modified conventional lithographic techniques and soft-lithographic methods, while bottom-up approaches include template-assisted patterning of HPs based on lithographically defined prepatterns and self-assembly. HP patterning is shown here to not only improve device performance, but also to reveal the unprecedented functionality of HPs, leading to new research areas that utilize their novel photophysical properties.     

Operational tactics and tenets of a new manufacturing paradigm ‘instant customerisation’

An increasing trend towards a new manufacturing paradigm of ‘instant customerisation’ is occurring due to customers demanding their orders for customised products or services be fulfilled under zero customer lead time and at a price near the price of mass production. By means of the manufacturing paradigm innovation model (MPIM), the operational tactics, operational tenets, and an operational framework of instant customerisation are elicited and described. An important new tactic within all the operational tactics is finalise-to-individual demand forecasting, meaning end products are finalised according to the forecasting of individual customer's needs. The knowledge and information, which is the primary input of individual demand forecasting, is gleaned and understood by total customer participation and value fusion. Total customer participation and value fusion are developed through customer participation and customer relationship management, which have critical distinctions. The operational framework, which is termed manufacturing paradigm tree of instant customerisation, can characterise the operational tenets, the relationships between different operational tactics and the objectives, and the relationships among different operational tactics. With the proposed framework, instant customerisation can be understood more effectively and holistically.     

A reliable decision support system for fresh food supply chain management

The paper proposes a decision support system (DSS) for the supply chain of packaged fresh and highly perishable products. The DSS combines a unique tool for sales forecasting with order planning which includes an individual model selection system equipped with ARIMA, ARIMAX and transfer function forecasting model families, the latter two accounting for the impact of prices. Forecasting model parameters are chosen via two alternative tuning algorithms: a two-step statistical analysis, and a sequential parameter optimisation framework for automatic parameter tuning. The DSS selects the model to apply according to user-defined performance criteria. Then, it considers sales forecasting as a proxy of expected demand and uses it as input for a multi-objective optimisation algorithm that defines a set of non-dominated order proposals with respect to outdating, shortage, freshness of products and residual stock. A set of real data and a benchmark – based on the methods already in use – are employed to evaluate the performance of the proposed DSS. The analysis of different configurations shows that the DSS is suitable for the problem under investigation; in particular, the DSS ensures acceptable forecasting errors and proper computational effort, providing order plans with associated satisfactory performances.     

Electrical nanowelding and bottom-up nano-construction together using nanoscale solder

A new bottom-up nanowelding technique enabling the welding of complex 3D nanoarchitectures assembled from individual building blocks using nanovolumes of metal solder is reported in this work. The building blocks of gold nanowires, (Co72Pt28/Pt)n multilayer nanowires, and nanosolder Sn99Au1 alloy nanowires were successfully fabricated by a template technique. Individual metallic nanowires dispersed on Si/SiO2(100 nm) wafers were manipulated and assembled together. Conductive nanostructures were then welded together by the new electrical nanowelding technique using nanovolumes of similar or dissimilar nanosolder. At the weld sites, nanoscale volumes of a chosen metal are deposited using nanosolder of a sacrificial nanowire, which ensures that the nanoobjects to be bonded retain their structural integrity. The whole nanowelding process is clean, controllable and reliable, and ensures both mechanically strong and electrically conductive contacts. The quality check of nanoweld achieve a resistance as low as 20 omega by using Sn99Au1 alloy solder. This technique should provide a promising way to conquer the challenge of the integration obstacle for bottom-up nanotechnology.     

Multivariate performance reliability prediction in real-time

This paper presents a technique for predicting system performance reliability in real-time considering multiple failure modes. The technique includes on-line multivariate monitoring and forecasting of selected performance measures and conditional performance reliability estimates. The performance measures across time are treated as a multivariate time series. A state–space approach is used to model the multivariate time series. Recursive forecasting is performed by adopting Kalman filtering. The predicted mean vectors and covariance matrix of performance measures are used for the assessment of system survival/reliability with respect to the conditional performance reliability. The technique and modeling protocol discussed in this paper provide a means to forecast and evaluate the performance of an individual system in a dynamic environment in real-time. The paper also presents an example to demonstrate the technique.     

Top-down and bottom-up mass spectrometric characterization of human myoglobin-centered free radicals induced by oxidative damage

In an effort to determine the utility of top-down mass spectrometric methodologies for the characterization of protein radical adducts, top-down approaches were investigated and compared to the traditional bottom-up approaches. Specifically, the nature of the radicals on human myoglobin induced by the addition of hydrogen peroxide and captured by the spin trap 5,5-dimethyl-1-pyrroline N-oxide (DMPO) was investigated. The most abundant ion observed in the electrospray mass spectrum of this reaction mixture corresponds in mass to the human myoglobin plus one DMPO molecule. In addition, a second ion of lower abundance is observed, which corresponds to a second DMPO molecule being trapped on myoglobin. Top-down analyses using Fourier transform ion cyclotron resonance mass spectrometry can be used to characterize proteins and, thus, were performed on several different charge-state ions of both the native and the mono-DMPO nitrone adduct of human myoglobin. Data produced from the top-down analyses are very complex yet information rich. In the case of DMPO-modified human myoglobin, the top-down data localized the DMPO spin trap to residues 97-110 of the myoglobin. The observation of the y43+5 fragment ion arising from C-terminal cleavage to the cysteine-110 residue in the MS/MS spectrum of DMPO-modified myoglobin and not in the unmodified myoglobin implicates a change to this residue, specifically, DMPO adduction. On the other hand, using the traditional bottom-up approach of peptide mapping and MS sequencing methodologies, two DMPO radical adducts on human myoglobin were identified, Cys-110 and Tyr-103. The bottom-up approach is more proven and robust than the top-down methodologies. Nonetheless, the bottom-up and top-down approaches to protein characterization are complementary rather than competitive approaches with each having its own utility.     

Nanoscale DNA tetrahedra improve biomolecular recognition on patterned surfaces

The bottom-up approach of DNA nano-biotechnology can create biomaterials with defined properties relevant for a wide range of applications. This report describes nanoscale DNA tetrahedra that are beneficial to the field of biosensing and the targeted immobilization of biochemical receptors on substrate surfaces. The DNA nanostructures act as immobilization agents that are able to present individual molecules at a defined nanoscale distance to the solvent thereby improving biomolecular recognition of analytes. The tetrahedral display devices are self-assembled from four oligonucleotides. Three of the four tetrahedron vertices are equipped with disulfide groups to enable oriented binding to gold surfaces. The fourth vertex at the top of the bound tetrahedron presents the biomolecular receptor to the solvent. In assays testing the molecular accessibility via DNA hybridization and protein capturing, tetrahedron-tethered receptors outperformed conventional immobilization approaches with regard to specificity and amount of captured polypeptide by a factor of up to seven. The bottom-up strategy of creating DNA tetrahedrons is also compatible with the top-down route of nanopatterning of inorganic substrates, as demonstrated by the specific coating of micro- to nanoscale gold squares amid surrounding blank or poly(ethylene glycol)-passivated glass surfaces. DNA tetrahedra can create biofunctionalized surfaces of rationally designed properties that are of relevance in analytical chemistry, cell biology, and single-molecule biophysics.     

Application of metal-coded affinity tags (MeCAT): absolute protein quantification with top-down and bottom-up workflows by metal-coded tagging

As the quantification of peptides and proteins extends from comparative analyses to the determination of actual amounts, methodologies for absolute protein quantification are desirable. Metal-coded affinity tags (MeCAT) are chemical labels for peptides and proteins with a lanthanide-bearing chelator as a core. This modification of analytes with non-naturally occurring heteroelements adds the analytical possibilities of inductively coupled plasma mass spectrometry (ICPMS) to quantitative proteomics. We here present the absolute quantification of recombinantly expressed aprotinin out of its host cell protein background using two independent MeCAT methodologies. A bottom-up strategy employs labeling of primary amino groups on peptide level. Synthetic peptides with a MeCAT label which are externally quantified by flow injection analysis (FIA)-ICPMS serve as internal standard in nanoHPLC-ESI-MS/MS. In the top-down approach, protein is labeled on cysteine residues and separated by two-dimensional gel electrophoresis. Flow injection analysis of dissolved gel spots by ICPMS yields the individual protein amount via its lanthanide label content. The enzymatic determination of the fusion protein via its β-galactosidase activity found 8.3 and 9.8 ng/μg (nanogram fusion protein per microgram sample) for batches 1 and 2, respectively. Using MeCAT values of 4.0 and 5.4 ng/μg are obtained for top-down analysis, while 14.5 and 15.9 ng/μg were found in the bottom-up analysis.     

Single molecule electronic devices

Single molecule electronic devices in which individual molecules are utilized as active electronic components constitute a promising approach for the ultimate miniaturization and integration of electronic devices in nanotechnology through the bottom-up strategy. Thus, the ability to understand, control, and exploit charge transport at the level of single molecules has become a long-standing desire of scientists and engineers from different disciplines for various potential device applications. Indeed, a study on charge transport through single molecules attached to metallic electrodes is a very challenging task, but rapid advances have been made in recent years. This review article focuses on experimental aspects of electronic devices made with single molecules, with a primary focus on the characterization and manipulation of charge transport in this regime.     

Comparison of Missing Data Imputation Methods in Time Series Forecasting

Time series forecasting has become an important aspect of data analysis and has many real-world applications. However, undesirable missing values are often encountered, which may adversely affect many forecasting tasks. In this study, we evaluate and compare the effects of imputation methods for estimating missing values in a time series. Our approach does not include a simulation to generate pseudo-missing data, but instead perform imputation on actual missing data and measure the performance of the forecasting model created therefrom. In an experiment, therefore, several time series forecasting models are trained using different training datasets prepared using each imputation method. Subsequently, the performance of the imputation methods is evaluated by comparing the accuracy of the forecasting models. The results obtained from a total of four experimental cases show that the -nearest neighbor technique is the most effective in reconstructing missing data and contributes positively to time series forecasting compared with other imputation methods.