Kubernetes provides powerful container orchestration features, like auto-scaling, which dynamically adjusts the scale of containerized applications. However, the default auto scaling mechanism in Kubernetes typically responds to workload changes only after they have occurred, which can result in resource mismatches. While there has been significant research into proactive auto-scaling approaches, most existing solutions struggle to adapt to the constantly evolving characteristics of prediction targets. This paper presents PAHPA, an intelligent, prediction-assisted approach aimed at enhancing Kubernetes' autoscaling capabilities, which integrates an prediction model, SLMD-LightGBM proposed in this paper, featuring a self updating mechanism for continuous prediction optimization, queueing theory analysis to improve resource allocation decisions, and a correction mechanism that refines prediction metrics using real-time data. A central insight of this paper is that while predictions are crucial, they should not be the sole basis for decisions and must be complemented by real-time monitoring data to ensure robust and adaptive autoscaling. Experimental results demonstrate that PAHPA significantly enhances system stability and reduces service latency, achieving a 9.5% lower Violation Rate (16.3%) than HPA (25.8%) and a 63% reduction in 99th percentile latency (880ms vs. 2400ms). This research highlights how combining predictive analytics with real-time monitoring can lead to more effective autoscaling strategies in cloud-native environments.

5G-WiFi heterogeneous network (HetNet) is considered to be an effective solution to the edge computing capability crisis, the introduction of WiFi networks can well share the burden of the 5G network. In this paper, based on 5G-WiFi HetNet, we formulate the optimization problem of joint task offloading and resource allocation to maximize the utility of uplink transmission rate and business cost while ensuring high reliability. In order to address the high nonlinearity caused by the binary variables and the fractional form. We utilize a fractional transform method and propose a two-layer iterative algorithm based on the penalty dual decomposition (PDD) approach to effectively solve this mixed-integer nonlinear programming (MINLP) problem. Simulation results verify the convergence of the proposed algorithm and show that it improves the utility of the system while ensuring high reliability.

Visual anomaly detection methods with localization are critically important for industrial manufacturing quality control. Because of the rarity of anomalies and the irregular variation of anomaly patterns, unsupervised methods have been widely explored. For the anomaly detection and localization tasks, the image reconstruction-based approaches have shown competitive performance. However, reconstruction results of those methods are coarse and visually blurred, which leads to a high rate of false detection and false pixel-level localization. To address those issues, we propose a framework called Dual-Path Re-construction Guided Segmentation Network (DRGS-Net), which determines the abnormal regions by segmenting the anomaly image with its reconstructed result as the reference template. DRGS-Net mainly consists of a novel dual-path reconstruction sub-network and a specifically designed anomaly segmentation sub-network. They are jointly trained end-to-end, with the former fusing texture repair and image reconstruction information to obtain fine-grained reconstruction results and the latter learning a decision boundary between normal and anomalous regions based on reconstruction results. On the standard benchmark dataset MVTecAD and an additional dataset DAGM, DRGS-Net shows competitive performance in image-level detection and achieves outstanding improvement in pixel-level localization. Further experiments with the few samples setting demonstrate that DRGS-Net retains strong performance with only few anomaly-free training images.

K-means algorithm is an effective clustering algorithm based on partition, which has been widely used for clustering analysis. However, there are two main problems for K-means algorithm: how to provide appropriate number of clusters and how to determine initial cluster centers automatically. Plenty of methods have been proposed to address the above problems. In our previous work, we proposed the hierarchical initialization approach to determine initial cluster centers, but we cannot provide the number of clusters automatically. In this paper, in order to determine the number of clusters automatically, we propose the Davies Bouldin Index (DBI) based hierarchical K-means (DHIKM) algorithm on the basis of our previous work. The proposed algorithm can integrate DBI metric into our hierarchical K-means algorithm and can determine the number of clusters with low time cost. Experiments on UCI datasets and synthetic data demonstrate the effectiveness and feasibility of the proposed algorithm.