Search results

This study aims to propose novel edge device-tailored federated learning algorithms of local classifiers (stochastic gradient descent, support vector machines), namely, FL-lSGD and FL-lSVM. These algorithms are designed to address the challenge of large-scale ImageNet classification.

The authors’ FL-lSGD and FL-lSVM trains in a parallel and incremental manner to build an ensemble local classifier on Raspberry Pis without requiring data exchange. The algorithms load small data blocks of the local training subset stored on the Raspberry Pi sequentially to train the local classifiers. The data block is split into k partitions using the k-means algorithm, and models are trained in parallel on each data partition to enable local data classification.

Empirical test results on the ImageNet data set show that the authors’ FL-lSGD and FL-lSVM algorithms with 4 Raspberry Pis (Quad core Cortex-A72, ARM v8, 64-bit SoC @ 1.5GHz, 4GB RAM) are faster than the state-of-the-art LIBLINEAR algorithm run on a PC (Intel(R) Core i7-4790 CPU, 3.6 GHz, 4 cores, 32GB RAM).

Efficiently addressing the challenge of large-scale ImageNet classification, the authors’ novel federated learning algorithms of local classifiers have been tailored to work on the Raspberry Pi. These algorithms can handle 1,281,167 images and 1,000 classes effectively.

This paper aims to establish a lattice Boltzmann (LB) method for solid-liquid phase transition (SLPT) from the pore scale to the representative elementary volume (REV) scale. By applying this method, detailed information about heat transfer and phase change processes within the pores can be obtained, while also enabling the calculation of larger-scale SLPT problems, such as shell-and-tube phase change heat storage systems.

Three-dimensional (3D) pore-scale enthalpy-based LB model is developed. The computational input parameters at the REV scale are derived from calculations at the pore scale, ensuring consistency between the two scales. The approaches to reconstruct the 3D porous structure and determine the REV of metal foam were discussed. The implementation of conjugate heat transfer between the solid matrix and the solid−liquid phase change material (SLPCM) for the proposed model is developed. A simple REV-scale LB model under the local thermal nonequilibrium condition is presented. The method of bridging the gap between the pore-scale and REV-scale enthalpy-based LB models by the REV is given.

This coupled method facilitates detailed simulations of flow, heat transfer and phase change within pores. The approach holds promise for multiscale calculations in latent heat storage devices with porous structures. The SLPT of the heat sinks for electronic device thermal control was simulated as a case, demonstrating the efficiency of the present models in designing and optimizing SLPT devices.

A coupled pore-scale and REV-scale LB method as a numerical tool for investigating phase change in porous materials was developed. This innovative approach allows for the capture of details within pores while addressing computations over a large domain. The LB method for simulating SLPT from the pore scale to the REV scale was given. The proposed method addresses the conjugate heat transfer between the SLPCM and the solid matrix in the enthalpy-based LB model.

The purpose of this paper is to develop a coupled lattice Boltzmann model for the simulation of the freezing process in unsaturated porous media.

In the developed model, the porous structure with complexity and disorder was generated by using a stochastic growth method, and then the Shan-Chen multiphase model and enthalpy-based phase change model were coupled by introducing a freezing interface force to describe the variation of phase interface. The pore size of porous media in freezing process was considered as an influential factor to phase transition temperature, and the variation of the interfacial force formed with phase change on the interface was described.

The larger porosity (0.2 and 0.8) will enlarge the unfrozen area from 42 mm to 70 mm, and the rest space of porous medium was occupied by the solid particles. The larger specific surface area (0.168 and 0.315) has a more fluctuated volume fraction distribution.

The concept of interfacial force was first introduced in the solid–liquid phase transition to describe the freezing process of frozen soil, enabling the formulation of a distribution equation based on enthalpy to depict the changes in the water film. The increased interfacial force serves to diminish ice formation and effectively absorb air during the freezing process. A greater surface area enhances the ability to counteract liquid migration.

The effective prediction of crude oil futures prices can provide a reference for relevant enterprises to make production plans and investment decisions. To the nonlinearity, high volatility and uncertainty of the crude oil futures price, a matrixed nonlinear exponential grey Bernoulli model combined with an exponential accumulation generating operator (MNEGBM(1,1)) is proposed in this paper.

First, the original sequence is processed by the exponential accumulation generating operator to weaken its volatility. The nonlinear grey Bernoulli and exponential function models are combined to fit the preprocessed sequence. Then, the parameters in MNEGBM(1,1) are matrixed, so the ternary interval number sequence can be modeled directly. Marine Predators Algorithm (MPA) is chosen to optimize the nonlinear parameters. Finally, the Cramer rule is used to derive the time recursive formula.

The predictive effectiveness of the proposed model is verified by comparing it with five comparison models. Crude oil futures prices in Cushing, OK are predicted and analyzed from 2023/07 to 2023/12. The prediction results show it will gradually decrease over the next six months.

Crude oil futures prices are highly volatile in the short term. The use of grey model for short-term prediction is valuable for research. For the data characteristics of crude oil futures price, this study first proposes an improved model for interval number prediction of crude oil futures prices.

The visible time window (VTW) calculation of satellites to ground targets is significant for Earth observation satellites' operation management and control. With the improvement of satellite maneuvering capability and the complexity of on-orbit observation tasks, the traditional VTW calculation methods can no longer meet the demands of satellite operation management and control due to a large amount of calculation and low efficiency. The purpose of this study is to propose a fast VTW calculation method based on map segmentation named map segmentation method (MSM), to improve the calculation efficiency, and further solve this problem.

The main feature of the MSM method is to segment the map and subsatellite trajectories and traverse the subsatellite points within a specific range around the target, significantly reducing the search space and the amount of computation and improving computational efficiency.

Numerical simulations for two satellite orbits are implemented to verify the feasibility of the proposed VTW calculation method, and the traditional traversal method (TM) is also performed for comparative analysis. The results show that the proposed method can obtain the same VTW, using less calculation time than the TM. The computational efficiency is significantly improved, especially for many tasks. The calculation time of observing 500 targets is saved by more than 70%, indicating a broad application prospect.

This paper proposes an original VTW calculation method based on map segmentation to improve the calculation efficiency. The simulation scenarios are designed to verify the accuracy and effectiveness of the proposed method, and the observation targets are randomly distributed on the map. For comparative analysis, the TM is also performed under the same simulation conditions.

Promoting electric vehicles (EVs) is an effective way to achieve carbon neutrality. If EVs are widely adopted, this will undoubtedly be good for the environment. The purpose of this study is to analyze the impact of network externalities and subsidy on the strategies of manufacturer under a carbon neutrality constraint.

In this paper, the authors propose a game-theoretic framework in an EVs supply chain consisting of a government, a manufacturer and a group of consumers. The authors examine two subsidy options and explain the choice of optimal strategies for government and manufacturer.

First, the authors find that the both network externalities of charging stations and government subsidy can promote the EV market. Second, under a relaxed carbon neutrality constraint, even if the government’s purchase subsidy investment is larger than the carbon emission reduction technology subsidy investment, the purchase subsidy policy is still optimal. Third, under a strict carbon neutrality constraint, when the cost coefficient of carbon emission reduction and the effectiveness of carbon emission reduction technology are larger, social welfare will instead decrease with the increase of the effectiveness of emission reduction technology and then, the manufacturer’s investment in carbon emission reduction technology is lower. In the extended model, the authors find the effectiveness of carbon emission reduction technology can also promote the EV market and social welfare (or consumer surplus) is the same whatever the subsidy strategy.

The network externalities of charging stations and the subsidy effect of the government have a superimposition effect on the promotion of EVs. When the network effect of charging stations is relatively strong, government can withdraw from the subsidized market. When the network effect of charging stations is relatively weak, government can intervene appropriately.

Comparing previous studies, this study reveals the impact of government intervention, network effects and carbon neutrality constraints on the EV supply chain. From a sustainability perspective, these insights are compelling for both EV manufacturers and policymakers.

Current methods for flow field reconstruction mainly rely on data-driven algorithms which require an immense amount of experimental or field-measured data. Physics-informed neural network (PINN), which was proposed to encode physical laws into neural networks, is a less data-demanding approach for flow field reconstruction. However, when the fluid physics is complex, it is tricky to obtain accurate solutions under the PINN framework. This study aims to propose a physics-based data-driven approach for time-averaged flow field reconstruction which can overcome the hurdles of the above methods.

A multifidelity strategy leveraging PINN and a nonlinear information fusion (NIF) algorithm is proposed. Plentiful low-fidelity data are generated from the predictions of a PINN which is constructed purely using Reynold-averaged Navier–Stokes equations, while sparse high-fidelity data are obtained by field or experimental measurements. The NIF algorithm is performed to elicit a multifidelity model, which blends the nonlinear cross-correlation information between low- and high-fidelity data.

Two experimental cases are used to verify the capability and efficacy of the proposed strategy through comparison with other widely used strategies. It is revealed that the missing flow information within the whole computational domain can be favorably recovered by the proposed multifidelity strategy with use of sparse measurement/experimental data. The elicited multifidelity model inherits the underlying physics inherent in low-fidelity PINN predictions and rectifies the low-fidelity predictions over the whole computational domain. The proposed strategy is much superior to other contrastive strategies in terms of the accuracy of reconstruction.

In this study, a physics-informed data-driven strategy for time-averaged flow field reconstruction is proposed which extends the applicability of the PINN framework. In addition, embedding physical laws when training the multifidelity model leads to less data demand for model development compared to purely data-driven methods for flow field reconstruction.

Protection motivation theory (PMT) explains that the intention to cope with information security risks is based on informed threat and coping appraisals. However, people cannot always make appropriate assessments due to possible ignorance and cognitive biases. This study proposes a research model that introduces four antecedent factors from ignorance and bias perspectives into the PMT model and empirically tests this model with data from a survey of electronic waste (e-waste) handling.

The data collected from 356 Chinese samples are analyzed via structural equation modeling (SEM).

The results revealed that for threat appraisal, optimistic bias leads to a lower perception of risks. However, factual ignorance (lack of knowledge of risks) does not significantly affect the perceived threat. For coping appraisal, practical ignorance (lack of knowledge of coping with risks) leads to low response efficacy and self-efficacy and high perceptions of coping cost, but the illusion of control overestimates response efficacy and self-efficacy.

First, this study addresses a new type of information security problem in e-waste handling. Second, this study extends the PMT model by exploring the roles of ignorance and bias as antecedents. Finally, the authors reinvestigate the basic constructs of PMT to identify how rational threat and coping assessments affect user intentions to cope with data security risks.

This paper aims to propose an improved version of portfolio optimization model through the prediction of the future behavior of stock returns using a combined wavelet-based long short-term memory (LSTM).

First, data are gathered and divided into two parts, namely, “past data” and “real data.” In the second stage, the wavelet transform is proposed to decompose the stock closing price time series into a set of coefficients. The derived coefficients are taken as an input to the LSTM model to predict the stock closing price time series and the “future data” is created. In the third stage, the mean-variance portfolio optimization problem (MVPOP) has iteratively been run using the “past,” “future” and “real” data sets. The epsilon-constraint method is adapted to generate the Pareto front for all three runes of MVPOP.

The real daily stock closing price time series of six stocks from the FTSE 100 between January 1, 2000, and December 30, 2020, is used to check the applicability and efficacy of the proposed approach. The comparisons of “future,” “past” and “real” Pareto fronts showed that the “future” Pareto front is closer to the “real” Pareto front. This demonstrates the efficacy and applicability of proposed approach.

Most of the classic Markowitz-based portfolio optimization models used past information to estimate the associated parameters of the stocks. This study revealed that the prediction of the future behavior of stock returns using a combined wavelet-based LSTM improved the performance of the portfolio.

In the context of fixed-wing aircraft wing assembly, there is a need for a rapid and precise measurement technique to determine the center distance between two double-hole components. This paper aims to propose an optical-based spatial point distance measurement technique using the spatial triangulation method. The purpose of this paper is to design a specialized measurement system, specifically a spherically mounted retroreflector nest (SMR nest), equipped with two laser displacement sensors and a rotary encoder as the core to achieve accurate distance measurements between the double holes.

To develop an efficient and accurate measurement system, the paper uses a combination of laser displacement sensors and a rotary encoder within the SMR nest. The system is designed, implemented and tested to meet the requirements of precise distance measurement. Software and hardware components have been developed and integrated for validation.

The optical-based distance measurement system achieves high precision at 0.04 mm and repeatability at 0.02 mm within a range of 412.084 mm to 1,590.591 mm. These results validate its suitability for efficient assembly processes, eliminating repetitive errors in aircraft wing assembly.

This paper proposes an optical-based spatial point distance measurement technique, as well as a unique design of a SMR nest and the introduction of two novel calibration techniques, all of which are validated by the developed software and hardware platform.