Search

Abstract Differential Evolution (DE) stands as a potent global optimization algorithm, renowned for its application in addressing a myriad of practical engineering issues. The efficacy of DE is profoundly influenced by its control parameters and mutation strategies. In light of this, we introduce a refined DE algorithm characterized by adaptive parameters and dual mutation strategies (APDSDE). APDSDE inaugurates an adaptive switching mechanism that alternates between two innovative mutation strategies: DE/current-to-pBest-w/1 and DE/current-to-Amean-w/1. Furthermore, a novel parameter adaptation technique rooted in cosine similarity is established, with the derivation of explicit calculation formulas for both the scaling factor weight and crossover rate weight. In pursuit of optimizing convergence speed whilst preserving population diversity, a sophisticated nonlinear population size reduction method is proposed. The robustness of each algorithm is rigorously evaluated against the CEC2017 benchmark functions, with empirical evidence underscoring the superior performance of APDSDE in comparison to a host of advanced DE variants.

Range-Doppler (RD) imaging has eclipsed synthetic aperture radar (SAR) imaging as the latest hotspot in the field of radar image object detection, owing to its low cost, high speed, and broad application scope. However, RD images are often of low quality due to the loss of effective features. Aiming at the problem of insufficient accuracy of the existing deep-learning-based sea surface RD image object detection, this article presents an improved YOLOv8 object detection algorithm for RD images based on feature enhancement and attention mechanism (YOLO-RDFEA). First, we have designed a feature extraction network DarknetSD with fewer parameters, to provide fine-grained information and compensate for the lack of abstract information. In addition, by introducing the coordinate attention (CA) mechanism in the feature-fusion stage, the model’s attention to spatial and channel features is improved. Moreover, the classification loss is improved using the slide loss function, which enhances the algorithm’s focus on the features of hard examples. Finally, comprehensive tests and evaluations are conducted using a self-built RD image dataset. Compared with the YOLOv8 baseline, the YOLO-RDFEA algorithm significantly reduced the misdetection of ships, its R was elevated by 17.9%. The all-category F1 score increased by 5.1% and mAP0.5 improved by 3.4%, which proved that the algorithm improves the detection and identification performance of all object categories. At the same time, the number of model parameters was reduced by 65.1%, which provides some basis for the algorithm deployment on the hardware platform.

Abstract Digital twin can be defined as a digital equivalent of an object of which it can mirror its behavior and status or virtual replicas of real physical entities in Cyberspace. To an extent, it also can simulate and predict the states of equipment or systems through smart algorithms and massive data. Hence, the digital twin is emerging used in intelligent manufacturing Systems in real-time and predicting system failure and also has introduced into a variety of traditional industries such as construction, Agriculture. Rare earth production is a typical process industry, and its Extraction Process enjoys the top priority in the industry. However, the extraction process is usually characterized by nonlinear behavior, large time delays, and strong coupling of various process variables. In case of failures happened in the process, the whole line would be shut down. Therefore, the digital twin is introduced into the design of process simulation to promote the efficiency and intelligent level of the Extraction Process. This paper proposes the techniques to build the rare earth digital twin such as soft measurement of component content, component content process simulation, control optimization strategy, and virtual workshop, etc. At the end, the validity of the model is verified, and a case study is conducted to verify the feasibility of the whole Digital twin framework.

Electroencephalogram(EEG) signals are generated from central nervous system which are difficult to disguise, leading to its popularity in emotion recognition. Recently,semi-supervisedlearning exhibits promisingemotion recognition performance by involving unlabeled EEG data into model training. However, if we first build a graph to characterize the sample similarities and then perform label propagation on this graph, these two steps cannotwell collaborate with each other. In this paper, we propose an OptimalGraph coupledSemi-Supervised Learning (OGSSL) model for EEG emotion recognition by unifying the adaptive graph learning and emotion recognition into a single objective. Besides, we improve the label indicator matrix of unlabeledsamples in order to directly obtain theiremotional states. Moreover, the key EEG frequency bands and brain regions in emotion expression are automatically recognized by the projectionmatrix of OGSSL. Experimental results on the SEED-IV data set demonstrate that 1) OGSSL achieves excellent average accuracies of 76.51%, 77.08% and 81.29% in three cross-sessionemotion recognition tasks, 2) OGSSL is competent for discriminative EEG feature selection in emotion recognition, and 3) the Gamma frequency band, the left/righttemporal, prefrontal,and (central) parietal lobes are identified to be more correlated with the occurrence of emotions.

Clustering ensemble technique has been shown to be effective in improving the accuracy and stability of single clustering algorithms. With the development of information technology, the amount of data, such as image, text and video, has increased rapidly. Efficiently clustering these large-scale datasets is a challenge. Clustering ensembles usually transform clustering results to a co-association matrix, and then to a graph-partition problem. These methods may suffer from information loss when computing the similarity among samples or base clusterings. Rich information between samples and base clusterings is ignored. Moreover, the results are not discrete. They need post-processing steps to obtain the final clustering result, which will deviate greatly from the real clustering result. To address this problem, we propose a co-clustering ensemble based on bilateral k-means (CEBKM) algorithm. Our algorithm can simultaneously cluster samples and base clusterings of a dataset, to fully exploit the potential information between the samples and the base clusterings. In addition, it can directly obtain the final clustering results without using other clustering algorithms. The proposed method, outperformed several state-of-the-art clustering ensemble methods in experiments conducted on real-world and toy datasets.

Feature selection and feature transformation are the two main approaches to reduce dimensionality, and they are often presented separately. In this study, a novel robust and efficient feature selection method, called FS-VLDA-L2,1 (feature selection based on variant of linear discriminant analysis and L2,1-norm), is proposed by combining a new variant of linear discriminant analysis and L2,1 sparsity regularization. Here, feature transformation and feature selection are integrated into a unified optimization objective. To obtain significant discriminative power between classes, all the data in the same class are expected to be regressed to a single vector, and the important task is to explore a transformation matrix such that the squared regression error is minimized. Therefore, we derive a new discriminant analysis from a novel view of least squares regression. In addition, we impose row sparsity on the transformation matrix through L2,1-norm regularized term to achieve feature selection. Consequently, the most discriminative features are selected, simultaneously eliminating the redundant ones. To address the L2,1-norm based optimization problem, we design a new efficient iterative re-weighted algorithm and prove its convergence. Extensive experimental results on four well-known datasets demonstrate the performance of our feature selection method.

Similarity search is a foundational task in time series data mining. Although there are many ways to measure the similarity of time series, a lot of evidence indicates that dynamic time warping (DTW) has the best robustness in many applications. Unfortunately, the expensive computational cost limits its application in large-scale databases. To speed up similarity search under DTW, we design a framework of two-stage similarity search for time series. In the first stage, we propose an improved lower bounding distance, which can be used to discard plenty of dissimilar series to get a set of candidate sequences. In the second stage, to efficiently get the retrieval result from the set of candidate sequences, we explore early abandoning strategy to avoid the full calculation of DTW. Extensive experiments are conducted on real-world data sets. The experimental results indicate that the proposed method can improve the retrieval efficiency of similarity search under DTW and guarantee no false dismissals.

The AdaBoost algorithm has the superiority of resisting overfitting. Understanding the mysteries of this phenomenon is a very fascinating fundamental theoretical problem. Many studies are devoted to explaining it from statistical view and margin theory. In this paper, this phenomenon is illustrated by the proposed AdaBoost+SVM algorithm from feature learning viewpoint, which clearly explains the resistance to overfitting of AdaBoost. Firstly, we adopt the AdaBoost algorithm to learn the base classifiers. Then, instead of directly combining the base classifiers, we regard them as features and input them to SVM classifier. With this, the new coefficient and bias can be obtained, which can be used to construct the final classifier. We explain the rationality of this and illustrate the theorem that when the dimension of these features increases, the performance of SVM would not be worse, which can explain the resistance to overfitting of AdaBoost.

Constructing a powerful graph that can effectively depict the intrinsic connection of data points is the critical step to make the graph-based semisupervised learning algorithms achieve promising performance. Among popular graph construction algorithms, low-rank representation (LRR) is a very competitive one that can simultaneously explore the global structure of data and recover the data from noisy environments. Therefore, the learned low-rank coefficient matrix in LRR can be used to construct the data affinity matrix. Consider the existing problems such as the following: (1) the essentially linear property of LRR makes it not appropriate to process the possible nonlinear structure of data and (2) learning performance can be greatly enhanced by exploring the structure information of data; we propose a new manifold kernelized low-rank representation (MKLRR) model that can perform LRR in the data manifold adaptive kernel space. Specifically, the manifold structure can be incorporated into the kernel space by using graph Laplacian and thus the underlying geometry of data is reflected by the wrapped kernel space. Experimental results of semisupervised image classification tasks show the effectiveness of MKLRR. For example, MKLRR can, respectively, obtain 96.13%, 98.09%, and 96.08% accuracies on ORL, Extended Yale B, and PIE data sets when given 5, 20, and 20 labeled face images per subject.