The rapid growth of interconnected systems, particularly within the Internet of Things (IoT), has significantly increased the volume of data exchanged across networks, making data security a critical challenge in environments with limited computational resources. Consequently, the design of lightweight encryption schemes that can simultaneously provide strong security and high efficiency has become increasingly important. In this article, a hybrid encryption scheme called SEC-Blowfish is proposed, which integrates Elliptic Curve Cryptography (ECC), chaotic mapping, and a dynamic S-Box generation mechanism. In the proposed approach, ECC is employed for secure key generation and exchange, while chaotic sequences and dynamically generated S-Boxes enhance the randomness of the encryption process and reduce predictable patterns in the ciphertext. This hybrid structure significantly strengthens the algorithm’s resistance against common cryptanalytic attacks, including statistical, linear, and differential analyses. To evaluate its effectiveness, the proposed algorithm is tested on text datasets of varying sizes, and the experimental results demonstrate that the method achieves strong data confidentiality while providing improved performance compared with several existing approaches. These findings indicate that SEC-Blowfish can serve as an efficient and reliable solution for securing data in resource-constrained environments, particularly in IoT systems and modern communication networks.

The traditional stethoscope auscultation used for the detection of lung diseases is hampered by its poor sensitivity, complexity of sound, and dependence on clinical competence. As a result, it frequently leads to diagnostic errors, treatment delays, and inaccessibility in settings with limited resources. These difficulties highlight the need for automated, inexpensive, and portable diagnostics. This paper presents the lightweight two-layer 1D-CNN (one-dimensional convolutional neural network) for lung illness classification called LDSC. The signals are processed using normalized MFCC (Mel-Frequency Cepstral Coefficients) features after being resampled to 4 kHz, enhanced (noise, time-stretch, pitch-shift), and divided into 3-second frames. According to simulation results, the suggested scheme achieves 98% accuracy, sensitivity, and specificity on the ICBHI (International Conference on Biomedical and Health Informatics) dataset and 99% on the KAUH (King Abdullah University Hospital) dataset, respectively.

Software-defined networks (SDN) have demonstrated considerable benefits in various practical domains by decoupling the control plane from the data plane, thus facilitating programmable network management. This paper presents a two-stage approach for solving the problem of controller placement called DEA-GAO. In the first stage, this strategy assumes the SDN network as a graph and using Data Envelopment Analysis (DEA) and relying on graph centrality metrics such as closeness centrality, betweenness centrality, and eigenvector centrality, calculates the efficiency of nodes to determine the optimal locations for deploying controllers. In the second stage, to allocate switches to controllers, the proposed strategy employs the Green Anaconda Optimization algorithm (GAO) to achieve an optimal allocation while considering network parameters such as average delay, load balancing, and reliability. Finally, to assess the efficacy of the proposed methodology, it is juxtaposed with three extant methods utilizing diverse datasets from the Internet Topology Zoo. The experimental findings indicate that the proposed approach significantly surpasses the existing methods, specifically the hybrid RDMCP-PSO algorithm, heuristic CPP algorithm and PSO algorithm in terms of both average delay (8.8%, 28.8% and 22.2% respectively) and controller utilization (1.5%, 7.3% and 32% respectively).

Software-Defined Networking (SDN) architectures, while inherently centralized, are prone to scalability issues, security breaches, and performance bottlenecks. The use of several controllers and the optimal distribution of traffic load across them is an effective countermeasure. This manuscript presents the AP-DQN framework, a methodology for identifying the optimal placement of controllers in SDN environments by combining adaptive clustering with deep reinforcement learning. Initially, the approach uses an improved version of the Affinity Propagation algorithm (AP), which consists of SDN peculiar criteria, i.e., end-to-end latency, traffic load, line running costs, and normalized security level, to cluster network switches, and thus identify possible zones for controller installation. Subsequently, a Deep Q-network (DQN) agent formalises the controller placement problem as a Markov decision process (MDP) and incrementally learns an optimal deployment policy that simultaneously optimises relevant performance metrics. The technique was tested against the Internet Topology Zoo dataset and benchmarked against known techniques, such as MODCEP, Multi-GA, and Random-CP. Simulation results show that AP-DQN can obtain 24 % improvement in load balancing, 25 % reduction in latency, 28 % reduction in link operational expenses, and 12 % improvement in normalized security level compared to comparative techniques.

Deep learning now underpins modern computer vision across various applications, including classification, detection, segmentation, retrieval, and generation. This survey first revisits foundational architectures—CNNs, early RNN-based vision pipelines, GANs, and the first Vision Transformers—then traces the shift to today’s practice (2023–2025): vision-language models (e.g., CLIP), large-scale self-supervised pretraining (e.g., DINOv2), promptable foundation segmentation (SAM), and diffusion-based generators (DDPM, latent diffusion). We distill the core design principles and training patterns behind these systems, analyze their strengths and limitations relative to classical approaches, and identify deployment-relevant concerns: compute efficiency, data governance, safety, and evaluation rigor. Looking ahead, we highlight parameter-efficient adaptation of vision foundation models (prompting/adapters), reliable few-shot protocols, and hardware-aware neural architecture search that meets latency and energy budgets. These priorities define concrete knowledge gaps that delineate the current research frontier. In this 2025 update, we introduce a concise survey of quantum–classical hybrids for vision, covering QCNNs, QViTs, and QGANs, and highlight their promise in small-data or tight-parameter regimes.

With the rapid expansion of cloud computing and the increasing complexity of network traffic, security has become one of the fundamental requirements of these infrastructures. In this context, intrusion detection plays an important role in identifying malicious activities and reducing the damage caused by cyber attacks. Despite recent advancements, many intrusion detection systems still face challenges such as the presence of redundant features, high correlation among data, and dependence on local optima, which can reduce their efficiency in real-world environments. In this research, a hybrid intelligent framework for intrusion detection in cloud environments is presented, based on feature dimensionality reduction using principal component analysis (PCA), classification with a multilayer perceptron (MLP) neural network, and evolutionary optimization of network parameters using a genetic algorithm (GA). In the proposed method, network traffic data are first normalized, then dimensionality reduction is applied to reduce computational complexity and eliminate redundancy and correlation among features. Subsequently, the neural network parameters are dynamically optimized by the genetic algorithm to prevent getting stuck in local optima. The experimental evaluation of the proposed framework was conducted on three standard and widely used datasets: NSL-KDD, CIC-IDS2017, and UNSW-NB15, which include various types of attacks such as DoS/DDoS, Probe, R2L, and U2R. The results show that the proposed method achieved an accuracy of 99.55

Vehicular Ad Hoc Networks (VANETs) are characterized by highly dynamic topologies, leading to frequent link breakages and challenging reliable routing. While clustering effectively mitigates topology instability, optimal Cluster Head (CH) selection and routing remain NP-hard problems. Despite various existing approaches, many current meta-heuristic routing protocols struggle to balance exploration and exploitation in highly dynamic VANET environments, often suffering from premature convergence and cluster instability under high mobility. To address these critical limitations, this paper proposes CRAHO, a novel hybrid meta-heuristic approach integrating the CSA and HHO for robust clustering-based routing in VANETs. Specifically, CSA is employed during the clustering phase to evaluate critical parameters—such as communication link quality and spatial distance—to form highly stable clusters. Subsequently, the routing phase leverages HHO based on distance metrics and node degrees to establish optimal, persistent inter-cluster paths. By formulating a comprehensive multi-objective fitness function, the CRAHO algorithm effectively coordinates exploration and exploitation. This approach guarantees QoS by minimizing routing overhead and end-to-end delay while maximizing the Packet Delivery Ratio (PDR). Simulation results demonstrate that the proposed CRAHO framework significantly outperforms benchmark routing protocols in maintaining network stability and optimizing data transmission in highly mobile vehicular environments. Specifically, compared to the baseline methods, CRAHO achieves improvements of 10.06

Let G be a locally compact group and S a weak "-closed translation invariant subspace of L-infinity(G). M.E.B. Bekka proved that S is the range of a projection on L-infinity(G) which commutes with translation if and only if S is the range of a projection on L(G) which commutes with convolution. Our first purpose in this paper is to generalize Bekka's results for a certain class of left Banach G-module. This result is used to show that G is amenable if and only if whenever X is a left Banach G-module and S is a weak"-closed right invariant subspace of X* which is complemented in X*, then S is the range of a projection on X* which commutes with convolution. Finally, we explore the link between the projections properties and amenability of group algebras.

Software-Defined Networks (SDNs) offer programmability and ease of management, but the architectural revolution poses very serious security threats to the conventional intrusion detection system. To overcome the aforementioned problems, this paper proposes Bedbug-HMM, a flow-level anomaly detection framework for SDN controllers that processes NetFlow measurements to identify traffic anomalies causing controller overload. The framework integrates NetFlow data extraction, PCA-enhanced preprocessing, Bedbug Metaheuristic Algorithm (BMA)-optimized Hidden Markov Models with quantile-discretized observation sequences, population-based ensemble scoring, MCC-optimized thresholds, and ECA-based mitigation policies for automated OpenFlow response.The experimental analysis performed on the NSL-KDD and UNSW-NB15 datasets proves the superior performance of the proposed approach over six baselines including clustering, MLP/GNN, Baum-Welch HMM, PSO-HMM, and GA-HMM (with + 5.9

Wireless edge networks must support delay-sensitive services under fluctuating traffic, interference, heterogeneous SLA priorities, and limited radio, computing, and memory resources. This paper proposes Serverless-PRONTO, an SLA-aware orchestration framework that jointly controls serverless function placement, task offloading, bandwidth allocation, CPU slicing, and warm-instance management. The system model captures uplink and downlink transmission, queueing, execution, cold-start initialization, and memory occupied by retained function instances. To solve the resulting dynamic, partially observable, mixed discrete-continuous problem, Serverless-PRONTO combines a physics-aware sparse graph encoder with multi-agent TD3 under centralized training and decentralized execution. The encoder represents inter-node coupling through channel, interference, distance, queue, resource, function-demand, and cold-start features, while top-$$\:K$$ attention limits signaling. An SLA-risk mechanism prioritizes requests according to urgency, queue state, service class, and cold-start probability. A sequential feasibility projection converts raw actor outputs into valid placement, bandwidth, CPU, and memory decisions. Simulations against four serverless and edge-orchestration baselines under varying traffic loads and network sizes show higher SLA satisfaction, lower end-to-end delay and cold-start ratio, and more stable scalability, demonstrating the benefit of jointly coordinating wireless resources and serverless runtime states.

Software-defined networks (SDN), owing to their centralized control architecture, provide high flexibility in network management, configuration, and monitoring; however, this architecture also introduces critical challenges related to scalability, performance bottlenecks, and quality of service (QoS) degradation under heavy and dynamic traffic conditions, particularly in large-scale and beyond 5G (B5G) networks with stringent real-time latency requirements. In such environments, the controller placement problem (CPP) becomes an inherently NP-hard multi-objective optimization task, where conventional sequential and heuristic methods struggle to explore the massive solution space within practical time constraints, thereby motivating the need for computationally scalable frameworks that can exploit parallel processing and high-performance computing (HPC) capabilities. To address these challenges, this paper proposes DeepWK-MSTC, an advanced multi-objective controller placement framework that integrates weighted Kmeans-based clustering with a deep learning-driven optimization mechanism. The proposed method leverages the inherent parallelism of Deep Monte Carlo Tree Search (Deep-MCTS) to enable concurrent rollouts and accelerated decision-making, while jointly optimizing three key objectives: minimizing average delay ratio (ADR), improving energy efficiency (EE), and balancing controller load under dynamic traffic patterns. By incorporating network topology characteristics and real-time traffic dynamics, DeepWK-MSTC efficiently avoids local optima and ensures stable optimization behavior. The effectiveness of the proposed framework is evaluated on six real-world network topologies from the Internet Topology Zoo, namely Aarnet, Chinanet, Deutsche Telekom, Colt, Cogent, and Tata, and compared against state-of-the-art baselines including ALO and ELA-RCP. Experimental results demonstrate that DeepWK-MSTC achieves an average reduction of 50.2% in ADR, an average energy saving of 26.45%, and a 24% decrease in maximum controller load, with an additional 11.5% relative ADR reduction compared specifically to ELA-RCP. Overall, by explicitly exploiting parallel optimization and HPC-oriented design principles, DeepWK-MSTC enhances resource utilization and ensures scalable, stable, and real-time-capable controller placement for large-scale SDN environments.

In this paper, we address the problem of fake news detection on online political news by proposing a semi‑supervised model that combines a binary multi‑objective grasshopper optimization algorithm (GOA) for feature selection with a self‑training classifier. The proposed GOA variant simultaneously minimizes the number of selected textual features and the classification error, and is specifically adapted to discrete feature spaces. After feature selection, a semi‑supervised self‑training scheme is applied to exploit both labeled and unlabeled news articles. The method is evaluated on two public political news datasets, BuzzFeed Political News (1627 news articles) and Random Political News (75 news articles). Experimental results show that the proposed scheme achieves up to 96.7

This study focuses on the important task of optimizing device clustering and assigning them to edge servers, while also implementing data redistribution in hierarchical semi-synchronous federated learning within the realm of advancing edge computing. Our research goal is to increase the performance and scalability of federated learning systems by improving resource allocation and data processing efficiency, which will in turn enhance edge computing frameworks. The current literature does not have thorough methods that can effectively combine model accuracy with optimal device clustering algorithms in hierarchical semi-synchronous federated learning, leading to below-par performance and inefficient use of resources. This difference highlights the need for creative measures that enhance not only model training accuracy but also the grouping of devices as opposed to current methods. The study utilizes a Graph Neural Network (GNN) to group IoT devices according to their hardware features and local datasets, then applies the K-means algorithm to create efficient device clusters. After that, Hybrid Data Redistribution is used to equalize local datasets in each cluster, and Proximal Policy resource allocation optimization algorithm is implemented to allocate devices to edge servers according to bandwidth usage, and energy consumption based on real-time updates, ultimately enabling hierarchical semi-synchronous federated learning to improve model training. The results show a 15% increase in clustering metrics compared to current algorithms, showcasing how our method improves device assignment and data redistribution in hierarchical semi-synchronous federated learning, addressing issues in model accuracy and resource optimization.

Cloud computing is growing exponentially, and data centers consume more and more energy. As a result, developing energy-efficient task scheduling algorithms has emerged as a prominent research problem and challenge. This paper presents a new method for priority-aware task scheduling in cloud data centers using the Hyper-Heuristic Bacterial Foraging Optimization (BFO-HH) algorithm. In addition, it introduces a new approach to dynamically selecting and combining 4 low-level heuristics (Task Selection, Virtual Machine Migration, Load Balancing, Resource Consolidation) to minimize operational cost, reduce energy consumption, and improve Quality of Service (QoS). All experiments were performed using the CloudSim 3.0.3 toolkit, over heterogeneous synthetic workloads comprising between 20 and 200 cloudlets. The performance of BFO-HH is compared with four well-known metaheuristic algorithms: Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Ant Colony Optimization (ACO), and Artificial Bee Colony (ABC). The experimental results demonstrate that, under the tested CloudSim environment and for workloads ranging from 20 to 200 tasks, BFO-HH consistently outperforms all comparative algorithms across multiple metrics. For example, for 200 tasks, BFO-HH exhibits 9.9% less energy consumption, 9.3% achieves shorter makespan, reduces Service Level Agreement (SLA) violations by 28%, increases resource utilization by 7%, and reduces operational cost by 14%. against the state-of-the-art base algorithms. Albeit such improvements are statistically significant, as evidenced by standard deviations and a 95% confidence interval.

Vehicle ad hoc networks consist of a number of nodes, each equipped with wireless communication equipment. In these networks, the destination for some data is all the vehicles present in the network which is named data dissemination. Due to the rapid changes in the topology of these networks, the dissemination and delivery of messages to all vehicles in inter-vehicle networks is considered a significant challenge. Various methods have been proposed to overcome these challenges. Among the existing methods, clustering seems to be an appropriate approach because an entity called the cluster head is responsible for delivering packets to the nodes within the cluster. Several clustering methods have been introduced for such networks, based on metaheuristic algorithms and machine learning. However, considering the dynamic nature of vehicle networks, there is practically no time available for data collection and the execution of these algorithms. It appears that utilizing inherent information, such as social behaviors and characteristics that do not require data collection, can be suitable for clustering vehicles and selecting the cluster head. In this paper, a method based on social features is proposed, known as social clustering-based data dissemination. In this method, initially, a number of nodes are selected as cluster heads based on their social characteristics, and then other nodes connect to the cluster heads based on speed and degree of the clusters. Simulation results show that using social behaviors in clustering improves packet delivery by 15

Due to the widespread use of computer networks and the Internet, the number of intruders increases annually, and the integration, security, and access to digital sources are faced with constant threats. Thus, the importance of maintaining security in such environments and the need to design a defense system to discover different threats have made scholars conduct studies and present modern and efficient Intrusion Detection Systems (IDSs). Such systems work with data features and determine traffic conditions by analysing these features. The features present in each dataset would determine the type of traffic. The large number of features in the problem environment and the unpredictable behavior of the network have made intrusion detection the central problem in the security of computer networks. In addition, the presence of unnecessary features in large numbers has made the feature selection problem an essential one in the IDS. This paper proposes an IDS based on the Multi-Objective Farmland Fertility (MOFF) algorithm to perform feature selection on the intrusion detection dataset. In addition, the Support Vector Machine (SVM), K-Nearest Neighbor (KNN), and Decision Tree (DT) have been used to estimate the ability of the selected features to make accurate predictions of attacks. Simulation results have shown that the proposed method reduces the time required for intrusion detection by reducing the number of features from 41 to 12, from 41 to 13, and from 48 to 12 on the KDD cup99, NSL-KDD, and UNSW-NB15 datasets, respectively. In addition, the classification’s accuracy values were obtained as 99.705%, 99.32%, and 99.20%, respectively, indicating an acceptable performance level. In addition, analyzing the calculation complexities shows that the proposed method is better than similar approaches.

In this study, a novel hybrid algorithm named the Reinforced Zebra Optimization Algorithm (RZOA) is proposed to solve the Traveling Salesman Problem (TSP). Initially, a discrete version of the Zebra Optimization Algorithm, referred to as DZOA, was developed, in which the continuous relationships of the original ZOA were transformed into a set of discrete operators to effectively update the routes. Subsequently, to enhance the intelligence of the operator selection process and to maintain a dynamic balance between exploration and exploitation, a mechanism based on Deep Reinforcement Learning was designed. Within the framework of the proposed RZOA, each zebra agent adaptively and self-learnedly makes decisions using a Deep Q-Network (DQN). Each agent observes both its own state and the overall population state, employs the DQN to select the optimal operator, and improves its decision-making policy dynamically through an experience replay memory. The state-action-reward structure was carefully designed to ensure an intelligent and balanced learning behavior between local and global search processes. The results of numerical experiments conducted on 42 standard benchmark datasets from the TSPLIB repository demonstrate that the proposed RZOA exhibits significant superiority over the compared algorithms in terms of solution quality, convergence speed, and performance stability. In particular, the proposed method achieves near-optimal performance with an average Percentage Deviation of the Best solution (PDB) below 1% and an average Percentage Deviation of the Average solution (PDA) typically below 0.5% for small and medium-scale instances, while maintaining PDA values within 4-5% for large-scale problems. Moreover, RZOA outperforms competing algorithms in the majority of benchmark cases, demonstrating superior robustness and consistency. Furthermore, the Friedman and Wilcoxon statistical analyses confirm this superiority at a 95% confidence level. Overall, by integrating the global search capability of the Zebra Optimization Algorithm with the adaptive decision-making power of the DQN, the proposed RZOA provides a novel, intelligent, and efficient approach for solving complex combinatorial optimization problems.

Renewable Energy Systems (RES) have become necessary with the growing need for sustainable energy globally; however, RES faces several problems such as power outages, instability and economic loss. These challenges are structured into three specific technical hurdles: (i) intermittency of wind and solar energy, (ii) forecasting uncertainty, and (iii) grid integration issues. In this context, recent technological trends in short-term wind power generation forecasting have increasingly moved towards hybridizing deep learning models with optimization algorithms to mitigate resource uncertainty. This study finds that Artificial Intelligence (AI) is one of the best solutions for overcoming these problems. To provide conceptual clarity, this research distinguishes AI as the overarching intelligent framework, while Machine Learning (ML) enables algorithmic learning of relationships from data, and Deep Learning (DL) utilizes hierarchical architectures for capturing complex non-linear temporal patterns. Methods such as Support Vector Machines (SVM) and Random Forest (RF) are applied alongside DL for predicting the amount of energy produced from RES and scheduling maintenance activities. To address existing research gaps, this research presents an innovative approach through combining a unique four-stage scientific data cleaning method with a two-phase AI stabilization system. The system uses the Markowitz Model as the first stage of determining the optimal hybrid ratio of wind and solar resources in addition to employing an energy-constrained battery smoothing process to reduce residual fluctuations. Results indicate that AI-based predictive maintenance can decrease the operational expenditure (OPEX) of RES up to 30.0%. Furthermore, the Long Short-Term Memory (LSTM) model demonstrated superior predictive power, achieving a Coefficient of Determination (R2) of up to 0.988, whereas a two-stage AI framework integrating Markowitz-based portfolio optimization and energy-constrained battery control achieves an average variance reduction of 95.4% in grid stability.

Wireless multi-hop networks facilitate communication by relaying packets from the source to the destination through relay nodes. These networks are often employed in long-distance communication scenarios, utilizing short-range transmissions to facilitate communication. One of the primary challenges in such networks is efficient routing and ensuring cooperation among relay nodes. Previous approaches have employed cooperative nodes to address this issue. However, most methods treat cooperative nodes as a static group across the entire network or select them based on attributes like proximity or movement history relative to the source node. Despite these efforts, the inherently dynamic and unstable nature of wireless multi-hop networks continues to pose challenges, particularly in terms of high latency and decreased packet delivery ratios. This paper proposes a novel method to optimize routing in wireless multi-hop networks by integrating game theory with the ant colony optimization (ACO) algorithm. The ACO algorithm identifies a set of cooperative nodes for each destination node, while the initial population of cooperative nodes is determined using the "Hunter-Stag" game theory model. In this game, nodes are treated as players, and the resulting strategy matrix serves as the input to the ACO algorithm. Virtual ants traverse the network, selecting the optimal route based on pheromone levels and criteria such as path quality and hop count. Simulation results demonstrate that, compared to existing methods, the proposed approach significantly reduces delay (25%), improves packet delivery ratio (15%), and decreases the number of hops required (20%). These improvements highlight the effectiveness of the combined game theory and ACO approach in enhancing the performance and reliability of wireless multi-hop networks.

The widespread use of Internet of Things (IoT) devices has brought along the importance of efficient resource management in fog computing systems in order to optimize the quality of experience of users. Task scheduling in fog computing systems acts as a bridge between tasks and the available resources, necessitating the use of advanced scheduling algorithms to optimize the parameters, such as deadline, response time, energy, load balancing, and cost. This paper describes the Novel Task Scheduling Approach for fog computing, named the Novel Task Scheduling Method (NTSM). In NTSM, the fog computing nodes are predicted, and the predicted values are partitioned into heavy and light groups for efficient allocation of tasks. This scalable approach optimizes the irregular cellular learning automata using the concept of the artificial rabbit algorithm for the allocation of tasks in the heavy groups, while the meta-heuristic algorithm optimizes the tasks in the light groups. The proposed approach has been proved effective in simulating the resource management process in fog computing, pertaining to the use of the IoT, in terms of optimized usage of energy, load balancing, deadline satisfaction, minimized response times, and reduced costs.