Multi-task and multi-view learning based on particle swarm optimization for short-term traffic forecasting

Design of Experiments vs. TOPSIS to Select Hyperparameters of Neural Attention Models in Time Series Prediction

Attention models are used in neural machine translation to overcome the challenges of classical encoder-decoder models. In the present research, design of experiments and TOPSIS methods are used to select hyperparameters of a neural attention model for time series prediction. The configurations selected by both methods are compared with out-of-sample data in time interval between January 2020 and April 2020 when global economies were significantly impacted due to Covid-19 pandemic. Results demonstrated that both selection methods outperformed each other in terms of different output features. On the other hand, our results with more than 95 % coefficient of determination and less than 0.23 % MAPE verified that neural attention models had strong capabilities in exchange rate prediction even in extraordinary situations in global economies.

Spatial Characteristics and Factor Analysis of Pollution Emission from Heavy-Duty Diesel Trucks in the Beijing-Tianjin-Hebei Region, China

Heavy-duty diesel trucks (HDDTs) contribute significantly to NO_X and particulate matter (PM) pollution. Although existing studies have emphasized that HDDTs play a dominant role in vehicular pollution, the spatial distribution pattern of HDDT emissions and their related socioeconomic factors are unclear. To fill this research gap, this study investigates the spatial distribution pattern and spatial autocorrelation characteristics of NO_X, PM, and SO₂ emissions from HDDTs in 200 districts and counties of the Beijing-Tianjin-Hebei (BTH) region. We used the spatial lag model to calculate the significances and directions of the pollutants from HDDTs and their related socioeconomic factors, namely, per capita GDP, population density, urbanization rate, and proportions of secondary and tertiary industries. Then, the geographical detector technique was applied to quantify the strengths of the significant socioeconomic factors of HDDT emissions. The results show that (1) NO_X, PM, and SO₂ pollutants emitted by HDDTs in the BTH region have spatial heterogeneity, i.e., low in the north and high in the east and south. (2) The pollutants from HDDTs in the BTH region have significant spatial autocorrelation characteristics. The spatial dependence effect was obvious; for every 1% increase in the HDDT emissions in the surrounding districts and counties, the local HDDT emissions increased by 0.39%. (3) Related factors analysis showed that the proportion of tertiary industries had a significant negative correlation, whereas the proportion of secondary industries and urbanization rate had significant positive correlations with HDDT emissions. Population density and per capita GDP did not pass the significance test. (4) The order of effect intensities of the significant socioeconomic factors was proportion of tertiary industry > proportion of secondary industry > urbanization rate. This study guides scientific decision making for pollution control of HDDTs in the BTH region.

Improving the Reliability of Photovoltaic and Wind Power Storage Systems Using Least Squares Support Vector Machine Optimized by Improved Chicken Swarm Algorithm

In photovoltaic and wind power storage systems, the reliability of the battery directly affects the overall reliability of the energy storage system. Failed batteries can seriously affect the stable operation of energy storage systems. This paper aims to improve the reliability of the storage systems by accurately predicting battery life and identifying failing batteries in time. The current prediction models mainly use artificial neural networks, Gaussian process regression and hybrid models. Although these models can achieve high prediction accuracy, the computational cost is high due to model complexity. Least squares support vector machine (LSSVM) is a computationally efficient alternative. Hence, this study combines the improved chicken swarm optimization algorithm (ICSO) and LSSVM into a hybrid ICSO-LSSVM model for the reliability of photovoltaic and wind power storage systems. The following are the contributions of this work. First, the optimal penalty parameter and kernel width are determined. Second, the chicken swarm optimization algorithm (CSO) is improved by introducing chaotic search behavior in the hen and an adaptive learning factor in the chicks. The performance of the ICSO algorithm is shown to be better than CSO using standard test problems. Third, the prediction accuracy of the three models is compared. For NMC1 battery, the predicted relative error of ICSO-LSSVM is 0.94%; for NMC2 battery, the relative error of ICSO-LSSVM is 1%. These findings show that the proposed model is suitable for predicting the failure of batteries in energy storage systems, which can improve preventive and predictive maintenance of such systems.