Search

This study investigated how social interaction among robotic agents changes dynamically depending on the individual belief of action intention. In a set of simulation studies, we examine dyadic imitative interactions of robots using a variational recurrent neural network model. The model is based on the free energy principle such that a pair of interacting robots find themselves in a loop, attempting to predict and infer each other's actions using active inference. We examined how regulating the complexity term to minimize free energy determines the dynamic characteristics of networks and interactions. When one robot trained with tighter regulation and another trained with looser regulation interact, the latter tends to lead the interaction by exerting stronger action intention, while the former tends to follow by adapting to its observations. The study confirms that the dyadic imitative interaction becomes successful by achieving a high synchronization rate when a leader and a follower are determined by developing action intentions with strong belief and weak belief, respectively., 8 pages, 5 figures

Long-term load forecasting (LTLF) is an essential process for strategical planning of the future needed extension in the power systems of any country. Besides, deep learning has become the heart of the machine learning paradigm, which is wildly used nowadays in many fields, and it also has become the current revolution in Artificial Intelligence (AI). In this paper, an optimized deep learning model based on Stacked Long Short-Term Memory Network (SLSTMN) is proposed. The architecture of the model is optimized to get the best configuration using Genetic Algorithm (GA). In addition, the hyperparameters of the model network are optimized using many deep learning techniques. During the optimization process, hundreds of model configurations are tested. The accuracy of this model is compared with many deep learning models and is compared against the related work in the field of LTLF. The dataset of the South Australia State (SA) power system is used to test the compared models. This data includes maximum daily load, daily maximum temperature, daily minimum temperature, weekday, the month, and holidays for 12 years from 2005 to 2016. SLSTMN achieves excellent accuracy and the lowest error value (almost 1%) when compared with other models on the same dataset and with related work models on different datasets.

In most law enforcement cases, judgment debtors have behaviors of evading execution in China, which seriously affects the authority of legal judgments and the judiciary’s credibility. Characteristics analysis of judgment debtors plays a vital role in finding out the concealed property and improving efficiency in handling law enforcement cases. Considering the advantages of hesitant fuzzy linguistic term sets (HFLTSs) representing the judgment debtors’ attributes and keeping all the original evaluation information on judgment debtors, we develop a hesitant fuzzy linguistic agglomerative hierarchical clustering (HFL-AHC) method to cluster judgment debtors and analyze the main characteristic of judgment debtors with concealing property. In some situations, the existing HFLTS distance cannot divide the judgment debtors. Therefore, we propose some new distance measures to classify the judgment debtors. The clustering results show that the judgment debtors who hide property have a poor evaluation of trading behavior, work, credibility, and consumption behavior.

Load forecasting is critical for power system operation and market planning. With the increased penetration of renewable energy and the massive consumption of electric energy, improving load forecasting accuracy has become a difficult task. Recently, it was demonstrated that deep learning models perform well for short-term load forecasting (STLF). However, prior research has demonstrated that the hybrid deep learning model outperforms the single model. We propose a hybrid neural network in this article that combines elements of a convolutional neural network (1D-CNN) and a long short memory network (LSTM) in novel ways. Multiple independent 1D-CNNs are used to extract load, calendar, and weather features from the proposed hybrid model, while LSTM is used to learn time patterns. This architecture is referred to as a CNN-LSTM network with multiple heads (MCNN-LSTM). To demonstrate the proposed hybrid deep learning model’s superior performance, the proposed method is applied to Ireland’s load data for single-step and multi-step load forecasting. In comparison to the widely used CNN-LSTM hybrid model, the proposed model improved single-step prediction by 16.73% and 24-step load prediction by 20.33%. Additionally, we use the Maine dataset to verify the proposed model’s generalizability.

Due to the rapid growth of network data, the authenticity and reliability of network information have become increasingly important and have presented challenges. Most of the methods for fake review detection start with textual features and behavioral features. However, they are time-consuming and easily detected by fraudulent users. Although most of the existing neural network-based methods address the problems presented by the complex semantics of reviews, they do not account for the implicit patterns among users, reviews, and products; additionally, they do not consider the usefulness of information regarding fine-grained aspects in identifying fake reviews. In this paper, we propose an attention-based multilevel interactive neural network model with aspect constraints that mines the multilevel implicit expression mode of reviews and integrates four dimensions, namely, users, review texts, products and fine-grained aspects, into review representations. We model the relationships between users and products and use these relationships as a regularization term to redefine the model’s objective function. The experimental results from three public datasets show that the model that we propose is superior to the state-of-the-art methods; thus showing the effectiveness and portability of our model.

Residential short-term energy consumption forecasting plays an essential role in modern decentralized power systems. The rise of innovative prediction methods able to handle the high volatility of users’ electrical load has posed the basis to accomplish this task. However these methods, which mostly rely on Artificial Neural Networks, require that a huge amount of users’ fine-grained sensitive consumption data are centrally collected to train a generalized forecasting model, with implications on privacy and scalability. This paper proposes an innovative architecture specifically designed to overcome this need. By exploiting Federated Learning and Edge Computing capabilities, many Long Short-Term Memory (LSTM) models are locally trained by different users based on their own historical energy consumption samples. Such models are then aggregated by a specific-purpose node to build a generalized model that is re-distributed for improved forecasting at the edge. For better forecasting, our proposed local training procedure takes as input relevant features related to calendar (i.e., hour, weekday and average consumption of previous days) and weather conditions (i.e., clustered apparent temperature), and the architecture can group users according to consumption similarities (using K-means) or socioeconomic affinities. We thoroughly evaluate the approach through simulations, showing that it can lead to similar forecasting performance than a state-of-the-art centralized solution in terms of Root Mean Square Error (RMSE), but with up to an order of magnitude lower training time and up to 50 times less exchanged data when samples are recorded at finer granularity than one hour. Nonetheless, it keeps sensitive data local and therefore guarantees users’ privacy.

As an essential component of Intelligent Transportation Systems (ITS), short-term traffic flow prediction is a key step to anticipate traffic congestion. Due to the availability of massive traffic data, data-driven methods with a variety of features have been applied widely to improve the traffic flow prediction. China has the longest total length of expressways in the world and there is significant information recorded when vehicles enter and exit the expressway. In this paper, we collect data at an expressway exit station in Shanghai, split the data according to its originating entry stations and predict the corresponding exit station traffic flow using the multi split traffic flows. First, the original records are collected, preprocessed, split, aggregated and normalized. Second, the Long Short-Term Memory (LSTM) model is applied to learn from the features of the overall flow and split traffic flows to predict the overall exit flow. The baselines are models which only overall flow information is considered. Compared with the baselines, in other models, the split flows according entry stations are also considered for prediction. Finally, the LSTM model is made comparison with the Convolutional LSTM(ConvLSTM), the K-Nearest Neighbor (KNN) model and the Support Vector Regression (SVR) model. When the information of overall flow and 6 split traffic flows is used and step is set to 11 (with 5 minute aggregation), the model prediction performs best. Compared with the best result of LSTM baseline model, the improvement of prediction accuracy is up to 5.48 percent by Mean Absolute Error (MAE).

Short-term prediction models for an ego-vehicle’s speed contributes to the improvement of vehicle safety, driveability, and fuel economy. To achieve these desired outcomes, an accurate forward speed prediction model and its successful implementation in a real system is a prerequisite. This paper compares six velocity prediction models based on two types of data-driven models, a Markov chain and a Recurrent Neural Network (RNN), by implementing them in an embedded system to evaluate their prediction accuracy and execution time. The inputs to each model are the driving information acquired on a specific route, such as internal vehicle information, relative speed and distance to the vehicle in the front of the ego-vehicle, and ego-vehicle’s location estimated by the GPS signal along with the B-spline roadway model. The proposed prediction models predict the velocity profile of the ego-vehicle up to the prediction horizon of 150 m. The parameters of the proposed models have been optimized using Hyper-parameter Optimization via Radial basis function and Dynamic coordinate search. By applying real driving data, the Markov chain-based models show slightly lower prediction accuracy but shorter execution time than those of the RNN-based models.

With the continuous development of global science and technology industry, the demand for power is increasing, so short-term power load forecasting is particularly important. At present, a large number of load forecasting models have been applied to short-term load forecasting, but most of them ignore the error accumulation in the iterative training process. To solve this problem, this article proposes a combined measurement model which combines stacked bidirectional gated recurrent unit (SBiGRU), complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) and error correction. In the first stage, SBiGRU model is established to study the time series characteristics of load series under the influence of temperature and holiday types. The error series generated in the prediction process of SBiGRU model reflects the error characteristics of load series; In the second stage, the error sequence is decomposed into several intrinsic mode functions (IMF) components and trend components by CEEMDAN algorithm. The SBiGRU model is established again for each component to learn and predict, and the predicted values of all components are reconstructed to get the error prediction results; Finally, sum the two-stage prediction results to correct the error. The accuracy of SBiGRU-CEEMDAN-SBiGRU combination model is evaluated by two public power load data. The experimental results show that the SBiGRU-CEEMDAN-SBiGRU combination model has better accuracy and stability than the traditional model.

Topic modeling is a probabilistic graphical model for discovering latent topics in text corpora by using multinomial distributions of topics over words. Topic labeling is used to assign meaningful labels for the discovered topics. In this paper, we present a new topic labeling method that uses automatic term recognition to discover and assign relevant labels for each topic, i.e., TLATR (Topic Labeling using Automatic Term Recognition). TLATR uses domain-specific multi-terms that appear in the set of documents belonging to a topic. The multi-term having the highest score as determined by the automatic term recognition algorithm is chosen as the label for that topic. To evaluate TLATR, we use two real, publicly available datasets that contain scientific articles’ abstracts. The topic label evaluation is done both automatically and using human annotators. For the automatic evaluation, we use Pointwise Mutual Information, Normalized Pointwise Mutual Information, and document similarity. For human evaluation, we employ the average rating method. Furthermore, we also evaluate the quality of the topic models using the Adjusted Rand Index. To prove that our novel method extracts relevant topic labels, we compare TLATR with two state-of-the-art methods, one supervised and one unsupervised, provided by the NETL Automatic Topic Labelling system. The experimental results show that our method outperforms or provides similar results with both NETL’s supervised and unsupervised approaches.

Because mechanical failures are accompanied by contingency and randomness, fault data is often difficult to obtain, and fault labels are also difficult to assign. The lack of data and fault labels have become important issues that restrict the development of fault diagnosis. The paper proposed a generalized Laplacian label prediction (GLLP) algorithm, which mainly uses the generalized Laplacian matrix and calculated a new locally smooth term. Therefore, data points with ambiguous and unclear labels will be assigned a small label value, while samples with more certain labels can get a more confident label value. The effectiveness of the method is verified on the public dataset and the real test rig dataset, and it is expected that this method can be extended to more complex mechanical system fault diagnosis.

Over the years, opposition-based Learning (OBL) technique has been proven to effectively enhance the convergence of meta-heuristic algorithms. The fact that OBL is able to give alternative candidate solutions in one or more opposite directions ensures good exploration and exploitation of the search space. In the last decade, many OBL techniques have been established in the literature including the Standard-OBL, General-OBL, Quasi Reflection-OBL, Centre-OBL and Optimal-OBL. Although proven useful, much existing adoption of OBL into meta-heuristic algorithms has been based on a single technique. If the search space contains many peaks with potentially many local optima, relying on a single OBL technique may not be sufficiently effective. In fact, if the peaks are close together, relying on a single OBL technique may not be able to prevent entrapment in local optima. Addressing this issue, assembling a sequence of OBL techniques into meta-heuristic algorithm can be useful to enhance the overall search performance. Based on a simple penalized and reward mechanism, the best performing OBL is rewarded to continue its execution in the next cycle, whilst poor performing one will miss cease its current turn. This paper presents a new adaptive approach of integrating more than one OBL techniques into Jaya Algorithm, termed OBL-JA. Unlike other adoptions of OBL which use one type of OBL, OBL-JA uses several OBLs and their selections will be based on each individual performance. Experimental results using the combinatorial testing problems as case study demonstrate that OBL-JA shows very competitive results against the existing works in term of the test suite size. The results also show that OBL-JA performs better than standard Jaya Algorithm in most of the tested cases due to its ability to adapt its behaviour based on the current performance feedback of the search process.

Classifying the cognitive levels of assessment questions according to Bloom’s taxonomy can help instructors design effective assessments that are well aligned with the intended learning outcomes. However, the classification process is time consuming and requires experience. Many studies have attempted to automate the process by utilizing different machine learning and text mining approaches, but none has examined the classification of Arabic questions. The purpose of this study is to examine this research gap and to introduce a new feature extraction method that would better suit Arabic questions and their unique characteristics. It also aims to provide Arab instructors with a tool that can help them automatically classify their assessment questions. To accomplish this purpose, the study developed a dataset of more than 600 Arabic assessment questions. It then proposed a modified term frequency-inverse document frequency (TF-IDF) method for extracting features from Arabic questions. Unlike the traditional TF-IDF, the proposed method was designed to take the nature of assessment questions into consideration. It was evaluated by comparing it to two methods that have been used for classifying English questions, i.e., the traditional TF-IDF and a modified TF-IDF method called term frequency part-of-speech-inverse document frequency (TFPOS-IDF). A t-test was utilized to examine whether the difference in performance between the three methods was statistically significant. The proposed method outperformed the two other methods. The overall accuracy, precision, and recall scored by the proposed method were significantly higher than those scored by the traditional TF-IDF and TFPOS-IDF methods. The evaluation results indicate the promising potential of the proposed method, which can be extended to other languages.

We propose a Generative Adversarial Network (GAN) optimized for noise reduction in CT-scans. The objective of CT scan denoising is to obtain higher quality imagery using a lower radiation exposure to the patient. Recent work in computer vision has shown that the use of Charbonnier distance as a term in the perceptual loss of a GAN can improve the performance of image reconstruction and video super-resolution. However, the use of a Charbonnier structural loss term has not yet been applied or evaluated for the purpose of CT scan denoising. Our proposed GAN makes use of a Wasserstein adversarial loss, a pretrained VGG19 perceptual loss, as well as a Charbonnier distance structural loss. We evaluate our approach using both applied Poisson noise distribution in order to simulate low-dose CT imagery, as well as using an anthropomorphic thoracic phantom at different exposure levels. Our evaluation criteria are Peek Signal to Noise (PSNR) as well as Structured Similarity (SSIM) of the denoised images, and we compare the results of our method versus recent state of the art deep denoising GANs. In addition, we report global noise through uniform soft tissue mediums. Our findings show that the incorporation of the Charbonnier Loss with the VGG-19 network improves the performance of the denoising as measured with the PSNR and SSIM, and that the method greatly reduces soft tissue noise to levels comparable to the NDCT scan.

In this article, we address the super-resolution problems, which estimate the high-resolution multispectral images from the multispectral Sentinel-2 (S2) images with different resolutions. Since S2 images can be naturally represented by tensors, we reformulate the degradation process as the tensor-based form. Based on the degradation mechanism, we build a tensor-based optimization model for S2 images super-resolution problem, which fully exploits intrinsic nonlocal spatial similarity and global spectral redundancy. Specifically, the model consists of the data fidelity term and the low-multirank regularizer tailored to thoroughly mining the inherent spatial-nonlocal and spectral redundancy. Then, we develop an efficient alternating direction method of multipliers algorithm with theoretically guaranteed convergence to tackle the resulting tensor optimization problem. Numerical experiments including simulated and real data demonstrate that our method outperforms the competing methods visually and qualitatively.

This tutorial guides the reader through the open literature on the aspect of Initial Cell Search (ICS) in both Code Division Multiple Access (CDMA) and Orthogonal Frequency Division Multiple Access (OFDMA) systems perspectives with a special focus on the DownLink (DL), where the ICS represents the very first operational stage of commercial mobile station modem’s always-on mobility. There is no comprehensive up-to-date tutorial of the ICSs encompassing 3G, 4G, and 5G systems. Hence, we review the ICS-related operations which have been proposed for Wideband CDMA (W-CDMA), Time Division-Synchronous CDMA (TD-SCDMA), Long Term Evolution Frequency Division Duplex (LTE-FDD), and Time Division LTE (TD-LTE) from the point of view of 3rd Generation Partnership Project (3GPP) standard and implementation. In the case of TD-SCDMA, there is no in-depth literature illustrating the details of ICS. Therefore, this tutorial aims to fill the related gap in the literature. As an evolution from the operational systems, we also shed light on the ICS being employed in New Radio (NR) system operating at sub-6 GHz and provide insight into interesting and challenging technical issues in millimeter-Wave spectrum bands, which is an ultimate foundation to be deployed in the full-scale NR system. Furthermore, we exemplify entire cell search operations over diverse spectrum bands and elucidate intelligent ways of designing multi-mode mobile station modem in a systematic manner. A taxonomy diagram also classifies three main targets constituting ICS procedures in 3G, 4G, and 5G systems. It is expected that this tutorial is beneficial not only for academic researchers but also for practical modem development engineers.

We propose a novel localization method using angle-of-arrival (AoA) measurements with two-step error variance-weighted least squares (TELS). The first step is to estimate a terminal location provisionally using least squares. The second step is to estimate the terminal location using weighted least squares, with the weights for each anchor and each evaluation-function term, calculated from the error variance based on the first step. The proposed method does not require previous information on the environment while achieving high performance. The simulation results indicate that a root mean square error (RMSE) of the proposed method is superior to that of the existing hybrid received signal strength (RSS)/AoA localization methods. When 11 anchors are deployed inside a cube with edge length 15 m, and the standard deviations of measurements are small, the RMSE of the proposed method reaches about 0.34 m. It is nearly equal to that of Cramer-Rao lower bound (CRLB) on AoA.

In order to meet the challenges brought by the high penetration of intermittent and fluctuating wind and solar power, a short-term optimal scheduling model for wind-solar-hydro hybrid generation system with cascade hydropower is established with the objective of minimizing the amount of abandoned wind, solar and hydro power and maximizing the stored energy of hydro stations. Cascade hydropower is considered to provide spinning reserve and regulation reserve to ensure the security of system. Mixed Integer Linear Programming (MILP) method is used for the short-term optimal schedule of wind-solar-hydro hybrid generation system. The case studies show that spinning reserve and regulation reserve are beneficial to the hybrid generation system, and verify the practical applicability of the proposed model.

Self-Regulated Learning (SRL) draws many scholars’ attention because of the fast proliferation of the quantity of information available to students and the advancements in educational technologies. This paper targets introducing a new approach to measure higher education students’ adoption of Self-Regulated Learning (SRL) strategies. The most common way to measure SRL is using very large or limited questionnaires, usually at the beginning of the term. The proposed approach is based on the administration of questions in small chunks over the term taking advantage of social media platforms (particularly Facebook). The paper describes the contents and procedures of the novel approach and compares it against the well-known Motivated Strategies for Learning Questionnaire (MSLQ). Firstly, the MSLQ was applied for initial data gathering (number of participants=344). Secondly, the new approach was administered via two Facebook groups voluntarily (number of participants=170). The study included students from diverse disciplines in different academic years. The findings show some differences between both instruments and reveal that the administration of questions in small chunks along the term leads to more reliable and internally consistent results (according to Cronbach’s coefficient values) and less affected by gender difference. Furthermore, SRL measurement in the novel approach takes place over an extended period, introducing some variations in the students’ answers. The results of a satisfaction questionnaire and interviews with non-active participants indicate the positive role of social media in enhancing students’ participation and emphasize the need for timely feedback to enhance the participation of students.

This study proposes a novel binary optimization method for solving unit commitment (UC) problems. Developed from moth-flame optimization (MFO), the proposed method is called binary alternative MFO (BAMFO). Originally, each MFO moth is restricted to fly toward a predetermined flame (solution), causing it to be trapped in the local optimum of that flame and reducing its performance. To address these drawbacks, BAMFO represents four alternative characteristics for moths to discover better flames all over the search space. The exploration and exploitation performances of BAMFO via these characteristics are balanced by a random parameter competing with the criteria of each characteristic. A repair strategy working with the priority list is created to adjust any unit status from the BAMFO simulation to satisfy all UC constraints. Then, Lambda-iteration method is adopted to solve economic dispatch problems. The proficiency of BAMFO is verified in various scales of unit systems for short-term operational scheduling. The results show that BAMFO yields better solutions than do other methods. The Wilcoxon signed rank test was conducted to prove the predominance of BAMFO.

Due to the uncertainty and fluctuation of distributed generation (DG) and load, the operation of active distribution network (ADN) is affected by multi-dimension factors which are described by massive operation scenarios. Efficient and accurate screening of severely restricted scenarios (SRSs) has become a new challenge in ADN planning. In this paper, a novel bi-level coordinated planning model which combines the short-time-scale operation problem with the long-time-scale planning problem is proposed. At the upper level, the demand response (DR) resource, an effective non-component planning resource characterized by low capacity price, high energy price, and short contract term, is co-optimized with the configuration of lines and energy storage systems (ESSs) to achieve the economic trade-off between the investment cost and the operation cost under SRSs. At the lower level, with the planning scheme obtained from the upper level, massive operation problems are optimized to minimize the daily operation cost; and the SRSs are provided to the upper level through a shadow-price-based scenario screening method, which simulates the planning information (i.e., the restricted degrees of operation scenarios) feedback process from ADN operators to ADN planners. Case studies on a 62-node distribution system in Jianshan New District, Zhejiang Province, China, illustrate the effectiveness of the proposed bi-level coordinated planning model considering DR resources and SRSs.

As the world witnesses population increase, the global power demand is increasing and the need for exploring other alternative clean and self-renewable sources of energy such as wind has become necessary. For optimal operation of the wind farms and stability of the grid, wind prediction ahead of time is of key importance. An accurate forecast of wind speed is often difficult due to the unpredictable nature of the wind. In this work, we utilized different machine learning models and proposed a hybrid machine learning approach. This approach combines 1D convolutional neural network (CNN) and bidirectional long short term memory (BLSTM) network for accurate prediction of short term wind prediction at different heights above ground level (AGL). The 1D CNN model extracts high-level features of the input wind speed data. The extracted features are then fed as input to the BLSTM network for wind speed prediction. The wind speed time series data used in this study are measured at 18, and 98 meters AGL. The study further presents a relationship between the utilized models and prediction accuracy at different heights. The forecasting performance of the models tends to increase as the height AGL increases. A real-world case study is implemented to demonstrate the effectiveness of the proposed CNN-BLSTM method in Saudi Arabia. The mean absolute error (MAE), mean squared error (MSE), root mean squared error (RMSE), and mean absolute percentage error (MAPE) are used as performance indices to evaluate the performance of the proposed CNN-BLTSM model. The corresponding results show that the proposed method outperforms other benchmark models.

There were necessary trajectory modifications of Cassini spacecraft during its last 14 years movement cycle of the interplanetary research project. In the scale 1.3 hour of signal propagation time and 1.4-billion-kilometer size of Earth-Cassini channel, complex event detection in the orbit modifications requires special investigation and analysis of the collected big data. The technologies for space exploration warrant a high standard of nuanced and detailed research. The Cassini mission has accumulated quite huge volumes of science records. This generated a curiosity derives mainly from a need to use machine learning to analyze deep space missions. For energy saving considerations, the communication between the Earth and Cassini was executed in non-periodic mode. This paper provides a sophisticated in-depth learning approach for detecting Cassini spacecraft trajectory modifications in post-processing mode. The proposed model utilizes the ability of Long Short Term Memory (LSTM) neural networks for drawing out useful data and learning the time series inner data pattern, along with the forcefulness of LSTM layers for distinguishing dependencies among the long-short term. Our research study exploited the statistical rates, Matthews correlation coefficient, and F1 score to evaluate our models. We carried out multiple tests and evaluated the provided approach against several advanced models. The preparatory analysis showed that exploiting the LSTM layer provides a notable boost in rising the detection process performance. The proposed model achieved a number of 232 trajectory modification detections with 99.98% accuracy among the last 13.35 years of the Cassini spacecraft life.

The study of climate change has become an important topic because of its negative impact on human life. The North-East African part lacks the studies for climate change detection, despite it being one of the most affected parts worldwide. The relationship between the emission of greenhouse gases (GHGs) and climate change is an important factor to understand. To investigate this linkage, we used machine-learning (ML) models based on essential climate variables (ECVs) to investigate the relationship between the GHGs and the rhythm of climate variable change. The article investigates how ML techniques can be applied to climatic data to build an ML model that is able to predict the state of climate variables for the short and long term. By selecting a candidate model, it will help in climate adaptation and mitigation, also determine at what level GHGs should be kept and their corresponding concentrations in order to avoid climate events and crises. The used models are long short-term memory, autoencoders, and convolutional neural network (CNN). Alternatively, the dataset has been selected from U.K. National Centre for Earth Observation and Copernicus Climate Change Services. We compared the performance of these techniques and the best candidate was the Head—CNN; based on performance metrics such as root-mean-squared-error: 5.378, 2.395, and 15.923, mean-absolute-error: 4.157, 1.928, and11.672, Pearson: 0.368, 0.649, and 0.291, and R2 coefficient: 0.607, 0.806, and 0.539 for the ECVs temperature, CO2, and CH4, respectively. We were able to link the GHG emission to ECVs with high accuracy based on the reading of this geographic area.

Due to the low dispatchability of wind power, the massive integration of this energy source in power systems requires short-term and very short-term wind power output forecasting models to be as efficient and stable as possible. A study is conducted in the present paper of potential improvements to the performance of artificial neural network (ANN) models in terms of efficiency and stability. Generally, current ANN models have been developed by considering exclusively the meteorological information of the wind farm reference station, in addition to selecting a fixed number of time periods prior to the forecasting. In this respect, new ANN models are proposed in this paper, which are developed by: varying the number of prior 1-h periods (periods prior to the forecasting hour) chosen for the input layer parameters; and/or incorporating in the input layer data from a second weather station in addition to the wind farm reference station. It has been found that the model performance is always improved when data from a second weather station are incorporated. The mean absolute relative error (MARE) of the new models is reduced by up to 7.5%. Furthermore, the longer the forecasting horizon, the greater the degree of improvement.

Evaluating the effect of blanket jamming is at the core of performance analysis and jamming/anti-jamming design for radar. Restricted to diverse jamming types and radar’s applications, it is challenging to put forward a unified framework for quantitative evaluation. To address this issue, we come up with a composite evaluation by combining the robust time-frequency analysis (RTFA) and peak to average power ratio (PAPR). In term of signal-level evaluation, RTFA is exploited to analyze the echoes directly, providing the time-frequency (TF) spectrum for calculating two-dimensional image entropy. For system/application-level evaluation, we derive the variation of signal to jamming ratio (SJR) in radar processing chain, and thus define the PAPR to associate SJR with target detection, a type of common and fundamental application that usually affects the other subsequent ones. To refine composite evaluation, we modify the traditional RTFA by leveraging joint sparse model with convolution framelets to improve TF concentration and to avoid the crossing terms; meanwhile, we derive the quantitative relationship between SJR and detection probability, leading to theoretical guarantee of PAPR for evaluation. Finally, the feasibility and the superiority of the proposed evaluation approach are validated in numerical experiments.

In this article, we comprehensively assessed the newly released long-term high-spatial-resolution Global LAnd Surface Satellite (GLASS) surface longwave (LW) radiation product using site measurements of LW fluxes. In total, three years of ground-measured LW fluxes (surface longwave upward radiation (LWUP), surface longwave downward radiation (LWDN), and surface longwave net radiation (LWNR) collected from 141 sites in six independent networks (AmeriFlux, AsiaFlux, BSRN, CEOP, HiWATER-MUSOEXE, and TIPEX-III) are used to evaluate the GLASS LW radiation product. These sites cover various land cover types, surface elevations, and climatic types. According to the evaluation results, the biases are -4.33, -3.77, and 0.70 W/m2 and the RMSEs are 18.15, 26.94, and 26.70 W/m2 for clear-sky LWUP, LWDN, and LWNR, respectively. The GLASS LW radiation product performs well in climate-change-sensitive areas such as poleward areas, semiarid areas, and the “third pole”, namely, the Tibetan Plateau. The accuracy of the GLASS LW product is higher or comparable to that of available LW products and studies but has a high-spatial-resolution of 1 km and a time span of 19 years. In conclusion, the overall accuracy of the clear-sky GLASS LW radiation product can satisfy the requirements of the hydrological, meteorological, and agricultural research communities on a global scale. We will continue to improve the retrieval algorithms and update the products accordingly.

Wind speed interval prediction is gaining importance in optimal planning and operation of power systems. However, the unpredictable characteristics of wind energy makes quality forecasting an arduous task. In this paper, we propose a novel hybrid model for wind speed interval prediction using an autoencoder and a bidirectional long short term memory neural network. The autoencoder initially extracts important unseen features from the wind speed data. The artificially generated features are utilized as input to the bidirectional long short term memory neural network to generate the prediction intervals. We also demonstrate that for time series prediction tasks, feature extraction through autoencoder is more effective than making deep residual networks. In our experiments which involve eight cases distributed among two wind fields, the proposed method is able to generate narrow prediction intervals with high prediction interval coverage and achieve an improvement of 39% in coverage width criterion over the traditional models.

Real-time voltage stability assessment (VSA) has long been an extensively research topic. In recent years, rapidly mounting deep learning methods have pushed online VSA to a new height that large amounts of learning algorithms are applied for VSA from the perspective of measurement data. Deep learning methods generally require a large dataset which contains measurements in both secure and insecure states, or even unstable state. However, in practice, the data of insecure or unstable state is very rare, as the power system should be guaranteed to operate far away from voltage collapse. Under this circumstance, this paper proposes an autoencoder based method which merely needs data of secure state to evaluate voltage stability of a power system. The principle of this method is that an autoencoder purely trained by secure data is expected to only create precise reconstruction for secure data, while it fails to rebuild data of insecure states. Thus, the residual of reconstruction is effective in indicating VSA. Besides, to develop a more accurate and robust algorithm, long short-term memory (LSTM) networks combined with fully-connected (FC) layers are used to build the autoencoder, and a moving strategy is introduced to bias the features of testing data toward the secure feature domain. Numerous experiments and comparison with traditional machine learning algorithms demonstrate the effectiveness and high accuracy of the proposed method.

Accurate and reliable traffic flow forecasting is of importance for urban planning and mitigation of traffic congestion, and it is also the basis for the deployment of intelligent traffic management systems. However, constructing a reasonable and robust forecasting model is a challenging task due to the uncertainties and nonlinear characteristics of traffic flow. Aiming at the nonlinear relationship affecting traffic flow forecasting effect, a PSO-ELM model based on particle swarm optimization is proposed for short-term traffic flow forecasting, which takes the advantages of particle swarm optimization to search global optimal solution and extreme learning machine to fast deal with the nonlinear relationship. The proposed model improves the accuracy of traffic flow forecasting. The traffic flow data from highways A1, A2, A4, A8 connecting to Amsterdam's ring road are employed for the case study. The RMSEs of PSO-ELM model are respectively 252.61, 173.75, 200.24, 146.05, while the MAPEs of PSO-ELM model are respectively 11.86%, 10.10%, 10.74%, 11.60%. The experimental results show that the performance of the proposal is significantly better than the performance of state-of-the-art models.

Most existing trackers perform well in the occurrence of short-term occlusions and appearance and illumination variations, but struggle with the challenges of long-term tracking which include heavy or long-term occlusions, and out-of-view objects. We propose a spatial-temporal aware long-term object tracking method in this paper, to manage the challenges of long term tracking. Firstly, we present a spatial-temporal aware correlation filter, which jointly models the spatial and temporal information of the target within the correlation filter framework to locate a visible target from frame to frame. Then, we design a tracking uncertainty detection mechanism to activate the re-detector in case of track failure. The mechanism relies on the variations of correlation response measured spatially, and the appearance similarity estimated temporally. Finally, we propose a spatial-temporal aware long-term re-detector to recover the target when it becomes visible. In the re-detection process, a large number of candidates are sampled, evaluated, and refined spatially to obtain accurate locations. Additionally, reliable memory retained through conservative template updating enables the recovery of the target by similarity matching. The spatial-temporal information is explicitly encoded in each component, which operates collaboratively to boost the overall performance. Extensive experimental results conducted on publicly available benchmark datasets demonstrate that the proposed method performs favorably when compared to other state-of-the-art trackers.

In this paper, a double time-delay feedback control of an optimal velocity model (OVM) is investigated. Double time-delay means that there exist two different state feedback control signals in the controlled OVM system, which are related to the velocity difference and the optimal velocity difference, respectively. Through linear stability analysis, the critical condition of Hopf bifurcation for the controlled OVM is derived. Utilizing the characteristics of Hopf bifurcation and the improved definite integral method, appropriate double time-delay feedback control strategy is designed in term of the number of unstable eigenvalues of the characteristic equation to suppress the stop-and-go waves generated by the uncontrolled OVM. Note that when the number of unstable eigenvalues is equal to zero, the controlled OVM is stable, otherwise, it is unstable. Numerical simulations are executed to validate the accuracy and feasibility of the design of double time-delay control strategy. Finally, case studies approximating the actual traffic situation are given, and the appropriate combination of control parameters is selected through the verified design steps. In addition, the measured data from NGSIM are also considered.

Due to the volatility and randomness of the photovoltaic power generation, it is difficult for traditional models to predict it accurately. To solve the problem, we established a model based on the self-attention mechanism and multi-task learning to predict the ultra-short-term photovoltaic power generation. First, we selected the data with the optimal timing length and input the data into the Encoder-Decoder network based on the self-attention. The validity of features extracted by the encoder was checked by the decoder. Then, we added a restriction to the middle layer of the Encoder-Decoder network to prevent the autoencoder from copying the input to the output mechanically. This condition is used to predict the photovoltaic power generation, so a multi-task learning model was established. Finally, to take full advantage of the features that are efficiently expressed and allow our main task, the prediction task, to learn some unique features autonomously, we proposed a step-by-step training method and have validated the effectiveness of this view in experiments. Through experimental contrast, it is found that compared with the Encoder-Decoder network based on CNN and LSTM, the performance of the proposed method has been increased by 14.82% and 8.09% respectively. The RMSE and MAE of the Encoder-Decoder model based on the self-attention mechanism using step-by-step training are 0.071 and 0.040 respectively.

The load resisting capacity of structural members will decrease when they are subjected to long-term sustained load. Such phenomenon is widely known as the duration of load effect, which is mainly caused by the damage accumulation in the material. The deterioration mechanism of the material is a typical stochastic process which is influenced by a large variety of parameters involving complex physical and chemical process. Although classical models have been proposed to evaluate the duration of load effect, it is nearly impossible to quantify the influence of various parameters and to achieve an accurate estimation. To optimize the combination of complexity and goodness-of-fit, a neural network model is proposed in this paper to evaluate the duration of load effect in wood structural members. Taking individual uncertainties into consideration, the proposed model treats the damage in wood as a Markov process and can estimate the residual strength distribution of the investigated wood structural members under long-term sustained load. The coefficient of determination reaches above 95% under sustained loading scenario, and it shows good adaptability across different wood properties. Moreover, the model can be adapted to continuously varied loading scenarios with a 98% coefficient of determination. This research aims to provide a useful and straightforward tool for accurately predicting the duration of load effect in wood structural members, and the proposed algorithm can be easily modified to deal with similar engineering problems for other construction materials.

Reconstruction of complex geological surface is widely used in oil and gas exploration, geological modeling, geological structure analysis, and other fields. It is an important basis for data visualization and visual analysis in these fields. The complexity of geological structures, the inaccuracy and sparsity of seismic interpretation data, and the lack of tectonic morphological information can lead to uncertainty in geological surface reconstruction. The existing geological surface uncertainty characterization and uncertain reconstruction methods have a shortcoming in balancing the interpolation error of high-confidence samples and model structure risk. Based on support vector regression (SVR), a method with confidence constraints for uncertainty characterization and the modeling of geological surfaces is proposed in this article. The proposed method minimizes the structural risk by adding a regularization term representing the model complexity, integrates high-confidence samples, such as drilling data, based on confidence constraints, and utilizes well path points by assigning appropriate inequality constraints to the corresponding prediction points. The results based on a real-world fault data set show that the uncertainty envelopes and fault realizations generated by the proposed method are constrained by well observations and well paths, effectively reducing the uncertainty.

This paper describes a scenario where D2D assisted cooperative relaying downlink transmission utilizing non-orthogonal multiple access (DC-NOMA) in the same time or frequency resource. Aiming to enhance the performance including ergodic sum rate (ESR) of the whole system and output block error rate (BER) of detection receiver. Firstly, we investigate the convexity of ESR maximization optimization problem and then prove its non-strictly concave property. Unlike existing works of optimizing the power coefficient only in one-stage, a jointly two-stage power allocation scheme based on iterative search is proposed for the formulated ESR maximizing problem. It is proved that the proposed algorithm converges to a stable state within limited numbers of iteration. Secondly, in order to improve the detection performance of the NOMA system downlink receiver, we propose a novel receiving detection algorithm based on the max-log-map algorithm and design a new receiver structure adapted to this algorithm. Finally we confirm the feasibility and verify the enhancement of the receiver design. The numerical results illustrate that: i) the jointly two-stage power allocation scheme can achieve up to 45.6% of performance gains in term of ESR in comparison of with the one-stage optimizing scheme; and ii) the delay and BER of the receiver both decrease significantly using novel receiver detection algorithm in the case of high-order signal modulation and multi-user multiplexing.

Reliable short-term prediction of available parking space (APS) is the basic theory of parking guidance information system (PGIS). Based on the Intelligent parking system at the Eastern New Town, Yinzhou District, Ningbo, China, this study collected the data of parking availability in the on-street parking areas. The variation characteristics of APS were investigated and analyzed at different spatial-temporal levels. Then the APS prediction models based on Gradient Boosting Decision Tree (GBDT) and Wavelet Neural Network (WNN) were proposed. Furthermore, an improved WNN algorithm with (WA) decomposition and Particle Swarm Optimization (PSO) were presented. The original time series was decomposed and reconstructed by wavelet analysis, and the WNN algorithm found the optimal threshold of initial weight through PSO. The result of GBDT (weekday: MSE = 27.37, SMSE = 0, TIME = 35min, weekend: MSE = 9.9, SMSE = 0, TIME = 35min) and WA-PSO-WNN (weekday: MSE = 14.93, SMSE = 1.88, TIME = 160.32s, weekend: MSE = 12.33, SMSE = 10.23, TIME = 160.95s) approximated the true value. But the prediction time of GBDT was too long to be applicable to the short-term prediction of APS in this paper. Compared with the methods of GBDT, WNN, and PSO-WNN, the WA-PSO-WNN algorithm performs much better. The average differences in MSE between WA-PSO-WNN and GBDT for weekday and weekend data are 45.45% and 58.76%, respectively, indicating that WA-PSO-WNN can increase the prediction accuracy of weekday and weekend data by an average of 45.45% and 58.76% compared with the GBDT model. Finally, the application prospects of short-term APS forecasting were also discussed in reducing cruising parking behavior, reducing illegal parking behavior and adjusting dynamic parking rates to verify the importance of APS short-term forecasting.

Short-term load forecasting is a critical task in the smart grid, which can be used to optimize power deployment and reduce power losses. Recurrent neural networks (RNNs) are the most popular deep learning models for short-term load forecasting. However, despite of achieving better forecasting accuracy than the traditional models, the performance of the existing RNN-based load forecasting approaches is still unsatisfactory for practical usage. Therefore, in this work, we have proposed input attention mechanism (IAM) and hidden connection mechanism (HCM) to greatly enhance the accuracy and efficiency of RNN-based load forecasting models. Specifically, we use IAM to assign the importance weights on input layers, which have better performances in both efficiency and accuracy than traditional attention mechanisms. To further enhance the models’ efficiency, HCM is then applied to utilize residual connections to enhance the model’s converging speed. We have applied both IAM and HCM on four state-of-the-art RNN implementations, and then conducted extensive experimental studies on two public datasets. Experimental results show that the proposed RNNs with IAM and HCM models achieve much better performances than the state-of-the-art baselines in both accuracy and efficiency. Ablation studies show that both IAM and HCM are essential to achieve such superior performances.

Wind power forecasting, to a certain extent, will transform the random fluctuation of wind power output into a known situation, which is one of the effective approaches to deal with large-scale wind power integrated into power grid. Due to the use of only historical data and the lack of new information, the accuracy of ultra-short-term wind power prediction (WPP) is still not satisfactory. Therefore, a combined prediction method based on the day-ahead Numerical Weather Prediction (NWP) location technology is proposed. Firstly, the time points with low forecasting accuracy of rolling WPP are approximately located by the NWP information and time windows, and then the hybrid approach combined with neural network and persistence method is presented to predict the future wind power output. The results of the case study show that compared with other classical prediction methods, this method can effectively improve the ultra-short-term prediction accuracy of wind power and verify the effectiveness of the proposed method.

The intermittent and uncertain nature of wind places a premium on accurate wind power forecasting for the reliable and efficient operation of power grids with large-scale wind power penetration. Herein, six-month-ahead wind power forecasting models were developed using tree-based learning algorithms. Three models were developed to investigate the impact of input data on forecasting accuracy. The first model was trained with the average and standard deviation of wind speed values measured at a height of 40 m with a 10-min sampling time. To evaluate the impact of sampling time on model performance, a second model was trained with wind speed values measured at a height of 40 m with 1-h, 12-h, and 24-h sampling times. To assess the effect of measuring height on model accuracy, the third model was trained with wind speed values measured at 40 m extrapolated from values measured at heights of 30 m and 10 m. Experiments revealed that using longer time intervals and height extrapolation leads to considerable accuracy degradation in forecasted models. Finally, to study the generalization ability of the forecasted models, they were tested against wind data measured at heights and locations different from what the models had been trained with. Simulation results substantiated that tree-based learning algorithms can be successfully adopted not only for long-term wind power forecasting, but for potential wind power forecasting at different heights and geographical locations.

Aspect-based sentiment analysis aims at identifying the sentiment polarity of specific aspect in the sentence. Previous work has realized the importance of the information interaction between aspect term and context. However, most existing information interaction methods are coarse-grained, which results in a certain loss of information. In addition, most methods ignore the role of position information in identifying the sentiment polarity of the aspect. To better address the two problems, we propose a novel approach, called hierarchical gated deep memory network with position-aware. Our approach has two characteristics: 1) it has fine-grained information interaction attention mechanism which models the word-level interaction between aspect and context. The sentence-to-aspect attention is used to capture the most indicative sentiment words in context. And the aspect-to-sentence attention is used to capture the most important word in the aspect term. 2) The position information is embedded as a feature in the sentence representation. Finally, we conduct sentiment classification comparative experiment on laptop and restaurant datasets. The experimental results show that our model achieves state-of-the-art performance on aspect-based sentiment analysis.

With the development of industrialization, fine particulate matter (PM2.5) severely chills the health of people. Studies have shown that the variation of PM2.5 concentration is related to the Global Navigation Satellite System (GNSS) tropospheric delay. Therefore, it is possible to use the widely distributed continuous operation reference station (CORS) to monitor and predict PM2.5 concentrations with high time resolution. In this paper, the zenith wet delay (ZWD) of five CORS located in Baoding, Hebei Province, China, from Sep 2014 to Feb 2015, is calculated firstly. Then, the correlation between PM2.5 and ZWD is investigated. Finally, the experimental data including PM2.5, ZWD, temperature, air pressure, and relative humidity data sampled in one hour are used to establish PM2.5 concentrations online prediction model based on support vector machine regression (SVMR) model with metabolic method. The experimental results show that in autumn and winter, the correlation coefficient between daily-mean PM2.5 and ZWD is mainly larger than 0.4, and the correlation coefficient between hourly mean PM2.5 and ZWD is mainly larger than 0.3. Meanwhile, in daily cycle analysis, air temperature, air pressure and relative humidity are related to PM2.5 concentration. Finally, Using SVMR model with metabolic method, by combining GNSS and meteorological factors, ideal short-term prediction accuracy is achieved, which shows that the use of GNSS and meteorological factors is potential in predicting PM2.5 concentration.

The existing forecast models for PM2.5 concentration can be classified into long term and short term models depending on whether the forecast is performed for the next few hours or days. However, short term forecast models feature narrow forecast time and are thus vulnerable in their sensitivity to soaring variations in air quality, defined as sudden events. The purpose of this work is to investigate the causes behind these sudden events. The PM2.5 data were obtained from monitoring devices deployed in Taichung as a part of the Airbox project. The data were fed into the current short-term forecast model to forecast air quality for the next hour. Event timing was detected by feeding the forecast result as an input to the sudden event detection model. We then combined the filtered timing with factors in environment and human activities. With the application of Hierarchy Clustering, the clustering result was analyzed to find the causes of sudden events. In the springtime and summertime, unexpected changes in rainfall and temperature were critical for forecast models. Moreover, unanticipated changes in the intensity of rainfall and wind are important in the autumn and winter. For human activities, crowds of commuters, tourists, and pilgrims also have influence on unusual air quality. By carefully considering the effects of sudden events, we believe that the response ability of short time forecast can improve significantly in the near future.

This study was conducted to develop a novel tracking control strategy for aeroengines with strong nonlinearity and uncertainty. Compared to existing robust gain-scheduling control strategies, the proposed control strategy has relatively low conservatism and can markedly improve engine performance. An improved on-board adaptive aeroengine model was established to estimate engine performance degradation and eliminate the degradation term contained in the perturbation block of the engine uncertain model in the design process. Robust controllers under engine normal and performance degradation states were designed at a set of operating points and scheduled according to relevant scheduling and health parameters. A desired robust gain-scheduling controller, which works based on performance degradation, can be precisely constructed via this approach. Simulation results are given to demonstrate the effectiveness of the proposed method, where the response speed of engine is improved by 38%.

Cloud computing has been widely applied in various fields with the development of big data and artificial intelligence. The associated resource demands exhibit characteristics such as diversity, large scale, burst and uncertainty. This paper analyzes these characteristics of cloud resource demands based on Alibaba cluster data, and proposes an adaptive short-term prediction algorithm for those demands. The proposed algorithm uses a principal component analysis method to extract the primary types of container demands from a time series of resource demands, and executes outlier detection and replacement to obtain a more stationary sequence. An adaptive short-term prediction strategy is proposed to adaptively select a higher-accuracy short-term prediction method to implement the prediction. Further, an error adjustment factor is proposed to reduce the prediction error. Thus, the short-term prediction accuracy of cloud resource demands is improved via outlier detection and replacement, an adaptive selection strategy and an error adjustment. We evaluated the effectiveness of these improvements, and compared our algorithm with existing algorithms in terms of effectiveness and time cost. The experimental results demonstrate that the proposed algorithm improves short-term prediction accuracy effectively.

Accurate and efficient ultra-short-term load forecasting is crucial for industrial power users to have stabilized and optimized operations. In this paper, we develop novel strategies for industrial power users to handle their challenges in ultra-short-term load forecasting. Firstly, this paper proposes a two-way Genetic Algorithm Back Propagation Neural Networks (GABPNN) missing data completion model to handle data loss, which is common power load data mining. A particle swarm optimization - supporting vector regression (PSO-SVR) algorithm is further used to integrate the two-way completion results with better accuracy. In addition, the paper introduces a combined ultra-short-term load forecasting model for industrial power users. The proposed model combines the Cubature Kalman filter (CKF) prediction model with good performance in nonlinear dynamic systems and the least square support vector machine (LS-SVM) prediction model with good performance in small-scale data prediction. The grey neural network is used to integrate the two algorithms, which further improves the accuracy of ultra-short-term load forecasting. Lastly, we test the proposed strategy in case study with real industrial data and demonstrate that the proposed model has a high degree of precision in load forecasting.

With the dramatic rise of mobile internet users and the administrative requirements of long-term data retention, telecom providers are facing increasingly challenging storage and retrieval issues of call detail records (CDRs). The existing storage system can only achieve the requirement of online query and offline analysis of the CDRs. However, to the best of our knowledge, few studies have focused on the topic of CDRs retrieval optimization with long-term storage. In order to improve the retrieval speed while ensuring a high compression ratio, in this paper we propose a novel hash storage scheme, termed dual-column bucketing (DCB), based on the Hive platform by making use of its Bucketing nature. Compared to the conventional scheme, the proposed DCB scheme can improve the performance both for CDRs compression and query. Second, similar storage scenarios such as storage of SMS, email and extended detail records (XDRs) are included in the optimization scope of the DCB. Experiments on real-world CDRs show that in contrast to the conventional scheme, the proposed DCB scheme can save the storage space by approximately 40%, reduces the amount of disk read to 2%, and improve the retrieval speed of known phone number queries by up to seven times.

Crustal movement and land subsidence will cause tower foundation settlement, which seriously affects the safety of transmission line operation. This paper proposed a short-term failure warning method for transmission tower under land subsidence condition based on finite element method, strain monitoring and time series prediction. By simulating the stress distribution of the tower in actual foundation settlement, the weak components of the settlement tower were founded. Combined with the finite element simulation results in various conditions, the reliability of simulation results is verified, and the strain variation range of the corresponding measuring point is obtained. By installing the real-time strain monitoring system and establishing the autoregressive integrated moving average (ARIMA) time series forecast model, warning of the failure of the tower can be sounded in short time according to the strain monitoring and forecasting results. The field results show that the tower is in good condition and with no risk in short term. The method proposed by this paper is economical and easy to operate, which can be used to conduct the emergency treatment of the dangerous tower under land subsidence condition.

In this paper, we consider the second-order consensus problem of hybrid multi-agent systems with unknown disturbances by using sliding mode control under the leader-follower network. First, the hybrid multi-agent system model with disturbances and nonlinear term is proposed, which is composed of continuous-time dynamic agents and discrete-time dynamic agents. Second, the definition of the second-order consensus of hybrid multi-agent system is given. Then, we assume that the interaction among all agents happens in sampling time and each continuous-time dynamic agent can observe its own states in real time. Based on the equivalent approaching law and the states information among agents, the sliding mode control protocols are designed to achieve the second-order consensus of the hybrid multi-agent system. Some sufficient conditions are given for solving the second-order consensus under the sliding mode control protocols. Finally, some simulations are also given to illustrate the validity of the proposed method.

The short-term scheduling schemes of cascaded hydropower plants are based on day-ahead hydrological forecasting information. Affected by the accuracy of prediction, real-term hydrological information can considerably vary from previously forecasted values, especially for the power load and local inflows. As a result, short-term scheduling schemes are difficult to apply directly in real-time scheduling. To solve this problem, a rolling optimal hourly operation model for interbasin cascaded hydropower plants is proposed to rapidly balance the load demand. The model is deemed an ultrashort-term model, and it has a short-cycle schedule between short-term and real-time schemes. The basic strategy of solving the model is as follows. First, the day-ahead power load and local inflow information are corrected using a statistical method based on the most recent real-time information. Second, the water delay time, defined as the time between when water is released from upstream reservoirs and arrives at downstream reservoirs, is updated based on the most recent real-time information. Third, a heuristic method is proposed to dynamically update the short-term scheduling scheme over the next few hours. The objective is to minimize the maximum relative deviation between the actual water level of a reservoir at the end of 24 hours and the day-ahead forecasted value. This approach can keep the reservoir water level at the end of 24 hours close to the forecasted value and improve the accuracy of estimating the initial water level the next day. The developed method is applied to solve an ultrashort-term scheduling problem involving the cascaded hydropower plants located in Yunnan Province, China. The results indicate that the proposed model can achieve seamless coupling between short-term scheduling and real-time scheduling. The proposed method can provide a scientific basis for the real-time dispatching of large-scale hydropower plants and improve the practicality of short-term dispatching schemes.