Concurrency and Computation-Practice & Experience最新文献

Memory models define the order in which accesses to shared memory in a concurrent system may be observed to occur. Such models are a necessity since program order is not a reliable indicator of execution order, due to microarchitectural features or compiler transformations. Concurrent programming, already a challenging task, is thus made even harder when weak memory effects must be addressed. A rigorous specification of weak memory models is therefore essential to make this problem tractable for developers of safety- and security-critical, low-level software. In this paper we survey the field of formalisations of weak memory models, including their specification, their effects on execution, and tools and inference systems for reasoning about code. To assist the discussion we also provide an introduction to two styles of formal representation found commonly in the literature (using a much simplified version of Intel's x86 as the example): a step-by-step construction of traces of the system (operational semantics); and with respect to relations between memory events (axiomatic semantics). The survey covers some long-standing hardware features that lead to observable weak behaviours, a description of historical developments in practice and in theory, an overview of computability and complexity results, and outlines current and future directions in the field.

We propose an incremental similarity-based label propagation algorithm (DLPA-S) for detecting dynamic community structures. As the network evolves, the method efficiently updates the communities over time via local label updates driven by changes in network topology—including edge and vertex additions or removals—and vertex similarity. This incremental approach significantly reduces computational cost while preserving accuracy in capturing community evolution. We evaluate DLPA-S using a comprehensive set of quality metrics that assess both the structural properties of the network and the agreement between detected communities and ground-truth partitions. Experiments are conducted on synthetic and real-world dynamic networks, varying key graph characteristics such as the number of vertices and the average degree, as well as across diverse community scenarios. The results show that DLPA-S consistently achieves stable and high-performing results, maintains high NMI and F1 scores, ensures strong internal connectivity, clear community separability, and avoids disconnected communities, while remaining computationally efficient.

To address the fraud in the rice market and the limitations of traditional traceability methods, this work proposes a novel framework integrating gas chromatography–mass spectrometry-based metabolomics with machine learning and heuristic feature extraction. A total of 190 japonica rice samples from six origins in Heilongjiang Province were analyzed, yielding 46 metabolite features. After mahalanobis distance quality control to eliminate outliers, 120 valid samples were retained and visualized via uniform manifold approximation and projection. Six representative machine learning algorithms were systematically evaluated, and genetic algorithm and simulated annealing were employed for feature optimization. The results show that the random forest algorithm combined with genetic algorithm achieved the highest performance (validation accuracy = 99.5%, AUC = 0.998, F-measure = 0.995), outperforming existing spectral and isotope-based methods. Twenty-eight key metabolites were identified, each closely linked to origin-specific environmental factors. Statistical tests confirmed significant performance differences between algorithms. This work provides a robust, interpretable, and cost-effective solution for rice origin traceability, with implications for food safety supervision and high-quality agricultural product authentication.

High-dimensional gene expression datasets pose significant challenges for cancer classification due to the presence of redundant and irrelevant features. To address this issue, we propose a hybrid framework that integrates the flower pollination algorithm (FPA) with support vector machines (SVM) for effective feature selection and classification. The FPA, inspired by the global and local pollination processes of flowering plants, is adapted into a binary variant using a sigmoid transfer function to select informative subsets of genes. The objective function balances classification accuracy with feature subset sparsity, thereby reducing dimensionality while preserving discriminative power. The selected gene subsets are subsequently evaluated using SVM, which provides robust classification in small-sample, high-dimensional scenarios. The proposed FPA-SVM framework was tested on multiple benchmark cancer datasets, including colon tumor, CNS, ALL-AML, breast cancer, lung cancer, ovarian cancer, lymphoma, MLL, and SRBCT. Experimental results demonstrate superior performance, with accuracy levels exceeding 98% for most binary-class datasets and competitive results for multiclass datasets, achieving up to 88.3% accuracy. These findings highlight the effectiveness of the proposed method in enhancing cancer classification, reducing dimensionality, and identifying potential biomarkers for precision medicine.

In reinforcement learning (RL), the assumption of fixed control frequency often leads to computational resource wastage and degraded policy performance, while traditional single-step temporal difference (TD) learning suffers from accumulated state-value estimation bias. This paper proposes the multi-state soft elastic actor-critic (MSSEAC) algorithm to address these issues: First, the paper introduces a temporal consumption penalty mechanism and reconstructs the actor network's dual-branch output structure to simultaneously generate control actions and time consumption estimates, enabling autonomous control frequency adjustment. Second, the multi-state temporal difference (MSTD) framework is developed to address the limitations of conventional single-step TD learning. Specifically, an innovative experience replay buffer management strategy is proposed, where historical actions are utilized to stabilize the learning process during initial training phases, with a gradual transition to policy-generated actions in later stages to enhance estimation accuracy. The multi-state-value estimation effectively mitigates the bias accumulation problem inherent in single-step TD methods through weighted fusion of return distributions from multiple future states. Code is available at: https://github.com/asdwqqqq/MSSEAC.git.

Virtual Network Embedding (VNE) plays a crucial role in optimizing physical network (PN) resource utilization in network virtualization and delivering service benefits such as isolation, cost efficiency, flexibility, security, and Quality of Service (QoS) to end users. Despite its importance, VNE faces significant challenges, such as assigning resources to Virtual Network Requests (VNRs) to drive lower energy consumption, which can unfavorably affect network performance. VNE constitutes two corresponding subproblems: virtual machine embedding and virtual link embedding, and both problems are treated as $��N��P�$ hard. In this context, minimizing energy consumption remains vital for SPs by effectively utilizing PN resources, as it not only increases the revenue-to-cost ratio but also enhances the acceptance of VNRs. This work introduces a novel heuristic framework called the Multi-Attributed Traffic Intensity Based Energy Aware Embedding for Online Virtual Network Requests (IViN) framework, designed to enhance the acceptance ratio while minimizing energy consumption. IViN considers a multi-attribute approach from system and network features in its heuristic ranking mechanism to rank virtual machines and servers during virtual machine embedding, followed by a virtual link assignment using the shortest path approach. These attributes play a crucial role in effectively capturing the dependencies between network elements. This helps IViN achieve energy-sensitive resource allocation and improves VNR acceptance and revenue-to-cost ratio. We validate the proposed approach by comparing it with existing methods through simulation experiments. The results show that IViN outperforms the baseline techniques by achieving improvements of 41%, 60%, and 34% in acceptance ratio, revenue-to-cost ratio, and energy consumption, respectively.

Surface defect detection is critical in industrial applications, especially for multi-scale small object detection under complex backgrounds, low resolution, and real-time constraints. Traditional YOLO-based detectors face three key limitations: single-path backbones struggle to balance fine-grained details and high-level semantics simultaneously, conventional feature pyramids cause small-object feature dilution through simple additive fusion, and detection heads lack explicit quality estimation for diverse defect morphologies. To address these challenges, we propose RFG-YOLO, a lightweight detection network. The proposed method introduces three key innovations: DPFNet employs dual-path parallel processing with PSHGNetv2 and YOLOv11 backbone to extract fine-grained details and semantic context simultaneously, eliminating the single-path bottleneck; MSWFPN utilizes weighted fusion modules including MFM, SCFM, and CSP-SMSFB to prevent small-object feature dilution through adaptive multi-scale integration; RQCD incorporates location quality estimation with RepConv to refine bounding box predictions for diverse defect shapes, improving localization precision across scales. Experiments on NEU-DET, GEAR-DET, and GC10-DET demonstrate that RFG-YOLO achieves mAP@50 improvements of 4.1%, 5.77%, and 4.1%, respectively, while reducing model parameters by 60.5% and computational cost by 50.8%. The model achieves a runtime speed of 156.0 frames per second on a GPU. These results validate that RFG-YOLO strikes an optimal balance between detection accuracy and computational efficiency, rendering it highly suitable for real-time defect detection in industrial settings.

The factorized sparse approximate inverse (FSAI) preconditioner has been proven to be effective in accelerating the convergence of iterative methods. Due to the high cost of constructing the FSAI preconditioner, accelerating it on graphics processing unit (GPU) has attracted considerable attention. However, despite the development of some existing FSAI preconditioning algorithms on GPU, their performance will significantly decrease when they encounter matrix types that are not suitable for them. This motivates us to investigate how to design an effective FSAI preconditioning algorithm on GPU. In this paper, we propose an adaptive FSAI preconditioning algorithm on GPU, called GFSAI-Adaptive, to address the above problem. In GFSAI-Adaptive, first, two adaptive thread allocation strategies are proposed for two special types of SPD matrices to ensure that the allocated threads can be fully utilized. Second, based on the proposed two thread allocation strategies, two FSAI kernels, called GFSAII and GFSAIII, are presented. Third, we construct a new graph convolutional network, and thus propose a search engine to select the optimal kernel from GFSAII and GFSAIII for matrices that do not belong to two special types based on it. Experimental results show that our proposed GFSAI-Adaptive is effective and outperforms a popular preconditioning algorithm in the public CUSPARSE library and a recent parallel static FSAI preconditioning algorithm on GPU.

The management of agricultural pests is crucial for the security of the food supply. The Brown Plant Hopper (BPH) harms rice fields and reduces productivity, which is a major concern. Accurate forecasting and early detection of BPH population dynamics based on environmental parameters are essential for the quick and efficient application of pest control strategies. This study investigates the use of machine learning methods like bagging, boosting, voting, and ensemble regression techniques to predict the population of BPH. The proposed ensemble stacking regression model uses Extreme gradient boosting (XGBoost), Random Forest (RF), and Extremely Randomized Trees (ERT) as base models with a meta-model Decision Tree (DT), and it is utilized to estimate the pest population. Several existing models are compared with the proposed model in terms of R-squared, Adjusted R-squared, RMSE as well as MSE. The outcomes show that these machine learning models are capable of accurately and successfully predicting the BPH population, which makes the proposed model an important tool for pest population prediction in rice farming.

Knee osteoarthritis (KOA) grading from x-ray images is important for supporting effective treatment planning. Yet it remains difficult due to the disease's complex presentation. Subtle anatomical changes add to the challenge. The subjectivity of manual evaluation further complicates the process. Such challenges highlight the importance of automated, objective, and reproducible computer-aided systems capable of leveraging complementary sources of information. In line with this, a joint fusion framework was developed to integrate optimized traditional image features with deep learning representations obtained from multiple pre-trained convolutional neural network models. Traditional features, including morphological, statistical, texture-based, and other clinically relevant descriptors, provide interpretable insights into bone structure and tissue characteristics. In parallel, deep features capture intricate spatial patterns and semantic details beyond the reach of manual modeling. For improved discrimination and efficiency, analysis of variance and linear discriminant analysis were used for selecting traditional features, while principal component analysis was applied to deep features to retain 85% variance. The two feature sets were combined using a Joint Fusion Type II approach, and class imbalance was mitigated through synthetic minority oversampling. A neural network was trained to capture interdependencies between these features. Experimental results on a benchmark KOA dataset indicated that fusion with VGG16 deep features achieved 85.39% accuracy, outperforming individual feature-based approaches. The framework maintained relatively high accuracy across all five KOA grades, including borderline cases, indicating its potential for consistent and clinically relevant KOA grading.