Georgios D. Karatzinis

The Cloud-Edge-IoT computing continuum requires mechanisms for grouping heterogeneous nodes into logical clusters capable of jointly hosting application workloads under multi-resource constraints. The core difficulty is selecting a subset of nodes that can collectively satisfy an application’s concurrent, multi-dimensional task demands, where aggregate resource sufficiency does not guarantee a feasible task-to-node assignment. Existing orchestration frameworks and simulators conflate node selection with task scheduling and cannot natively represent indivisible multi-resource joint feasibility under controlled demand stress. We address this gap with LCF-Bench, a reproducible benchmark for feasibility-aware logical cluster formation. It formalizes logical cluster formation as a feasibility-first node-selection problem, decoupled from runtime scheduling, and packages a parameterized instance generator, a Kubernetes-compatible export, and a standardized evaluation protocol spanning node scaling, task scaling, and demand-load scaling. Using the benchmark, we find that feasibility exhibits sharp phase transitions under increasing demand. The region of exact-optimization hardness is workload-dependent, concentrating near the feasibility transition under balanced and compute-heavy workloads yet largely vanishing under storage-dominated ones. Single-objective heuristics incur a heterogeneity-induced over-provisioning anomaly on skewed resource distributions. We report reference results for five representative methods: Round- Robin, First-fit, Best-fit, a Mixed-Integer Linear Program (MILP) exact reference, and the Kubernetes default scheduler. Best-fit offers the strongest practical trade-off, retaining full feasibility to the highest demand overload of any heuristic (roughly 80% in the reference run), while the MILP reference is optimal but intractable at scale. LCF-Bench provides a common, extensible reference point for the systematic evaluation of logical cluster formation methods under identical conditions.

Efficient hay bale detection and counting are essential tasks within modern precision agriculture, significantly impacting yield estimation, logistics, and sustainable resource management. To address current limitations in dataset quality and environmental representation, we introduce BaleUAVision , a comprehensive dataset consisting of 2,599 high-resolution RGB images, each containing numerous human-annotated hay bales. Captured by Unmanned Aerial Vehicles (UAVs) across 16 diverse agricultural fields in Northern Greece, the dataset includes varying flight altitudes (50–100 meters), diverse speeds (3.7–5 m/s), and overlapping strategies to ensure robust data representation. BaleUAVision provides rich annotations through polygon-based semantic segmentation in multiple formats (COCO, CSV, JSON, YOLO, segmentation masks) and high-quality orthomosaics for precise spatial analysis. Technical validation demonstrated the dataset’s effectiveness in training robust hay bale detection models using YOLOv11, achieving high precision and recall under varying geographic and altitude conditions. Specifically, the dataset supported effective generalization across geographically distinct areas (Xanthi and Drama regions) and varying altitudes, highlighting its utility in real-world UAV operations. The dataset and supplementary tools, scripts, and analyses are publicly available on Zenodo and GitHub respectively, following FAIR principles to support wide-reaching applicability within the research community.

Fuzzy Cognitive Maps (FCMs) have emerged as powerful tools for addressing diverse engineering challenges, leveraging their cognitive nature and ability to encapsulate causal relationships. This paper provides a comprehensive review of FCM applications across 15 engineering sub-domains, categorizing 80 studies by their learning family, task type, and case-specific application. We analyze the methodological advancements and practical implementations of FCMs, showcasing their strengths in areas such as decision-making, classification, time-series, diagnosis, and optimization. Qualitative criteria are systematically applied to classify FCM-based methodologies, highlighting trends, practical implications of varying complexity, and human intervention across task types and learning families. However, this study also identifies key limitations, including scalability challenges, reliance on expert knowledge, and sensitivity to data distribution shifts in real-world settings. To address these issues, we outline key areas and directions for future research focusing on adaptive learning mechanisms, hybrid methodologies, and scalable computational frameworks to enhance FCM performance in dynamic and evolving contexts. The findings of this review offer a structured roadmap for advancing FCM methodologies and broadening their application scope in both contemporary and emerging engineering domains.

The challenge of maintaining optimal comfort in residents while minimizing energy consumption has long been a focal point for researchers and practitioners. As technology advances, reinforcement learning (RL)—a branch of machine learning where algorithms learn by interacting with the environment—has emerged as a prominent solution to this challenge. However, the modern literature exhibits a plethora of RL methodologies, rendering the selection of the most suitable one a significant challenge. This work focuses on evaluating various RL methodologies for saving energy while maintaining adequate comfort levels in a residential setting. Five prominent RL algorithms—Proximal Policy Optimization (PPO), Deep Deterministic Policy Gradient (DDPG), Deep Q-Network (DQN), Advantage Actor-Critic (A2C), and Soft Actor-Critic (SAC)—are being thoroughly compared towards a baseline conventional control approach, exhibiting their potential to improve energy use while ensuring a comfortable living environment. The integrated comparison between the different RL methodologies emphasizes the subtle strengths and weaknesses of each algorithm, indicating that the best selection relies heavily on particular energy and comfort objectives.

As the field of fault diagnosis in electrical machines has significantly attracted the interest of the research community in recent years, several methods have arisen in the literature. Also, raw data signals can be acquired easily nowadays, and, thus, machine learning (ML) and deep learning (DL) are candidate tools for effective diagnosis. At the same time, a challenging task is to identify the presence and type of a bearing fault under noisy conditions, especially when relevant faults are at their incipient stage. Since, in real-world applications and especially in industrial processes, electrical machines operate in constantly noisy environments, a key to an effective approach lies in the preprocessing stage adopted. In this work, an evaluation study is conducted to find the most suitable signal preprocessing techniques and the most effective model for fault diagnosis of 16 conditions/classes, from a low-workload (computational burden) perspective using a well-known dataset. More specifically, the reliability and resiliency of conventional ML and DL models is investigated here, towards rolling bearing fault detection, simulating data that correspond to noisy industrial environments. Diverse preprocessing methods are applied in order to study the performance of different training methods from the feature extraction perspective. These feature extraction methods include statistical features in time-domain analysis (TDA); wavelet packet decomposition (WPD); continuous wavelet transform (CWT); and signal-to-image conversion (SIC), utilizing raw vibration signals acquired under varying load conditions. The noise effect is examined and thoroughly commented on. Finally, the paper provides accumulated usual practices in the sense of preferred preprocessing methods and training models under different load and noise conditions.

Fuzzy cognitive networks (FCNs) arose from traditional fuzzy cognitive maps (FCMs) to have the advantage of guaranteed convergence to equilibrium points, thus being more suitable than conventional FCMs for a variety of pattern recognition and system identification tasks. Moreover, recent developments led to FCNs with functional weights (FCNs-FW), as a significant FCNs enhancement in terms of storage requirements, efficiency and less human intervention requirements. In this paper we proceed further by introducing hybrid deep learning structures, interweaving FCNs-FW with well established deep neural network (DNN) representative structures and apply the new schemes on a variety of pattern recognition and time series prediction tasks. More specifically, after discussing general issues related to the construction of deep learning structures using FCNs-FW we present three hybrid models, which combine the FCN-FW with convolutional neural networks (CNNs), echo state networks (ESNs) and AutoEncoder (AE) schemes, respectively. The hybrid schemes are tested on diverse benchmark data sets and prove that FCN-FW based hybrid schemes perform equally well or better than state-of-the-art DNN-based schemes, paving thus the way for using cognitive networks to deep learning representative structures.