Lyamine Guezouli
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Lyamine Guezouli has a total of 63 co-authors. Here are the top co-authors by number of shared publications:
Kamel Barka
L@STIC Laboratory, Department of Computer Science, University of Batna 2 Algeria
7 shared publications- Djallel Eddine Boubiche
LEREESI (Laboratory of Renewable Energy, Energy Efficiency and Smart Systems) Laboratory, HNS-RE2SD (Higher National School of Renewable Energies, Environment and Sustainable Development), Batna 05000, Algeria
5 shared publications - Homero Toral-Cruz
Departamento de Ingeniería y Tecnología, Universidad Autónoma del Estado de Quintana Roo, Chetumal 77019, Mexico
5 shared publications - Rafael Martínez-Peláez
Departamento de Ingeniería de Sistemas y Computación, Universidad Católica del Norte, Antofagasta 1270709, Chile
5 shared publications - Kamel Barka
University of Batna 2
4 shared publications Souheila Bouam
University of Batna, LaSTIC Research Lab, 2 Batna, 05000, Algeria
4 shared publications
Latest Publications
- Open Access
Human Behavior in Cybersecurity from a Sociotechnical Perspective: A Literature Review of Theories and Dimensions
Human behavior in cybersecurity results from the interplay among individual appraisals, organizational conditions, technological factors, and adversarial contexts. This structured literature review investigates the functions of behavioral theories, their extensions and integrations, and the explanatory roles attributed to seven sociotechnical dimensions. We conducted four topic queries in the Web of Science Core Collection in January 2026 and updated the screening in October 2026, targeting journal articles published between 2015 and 2025. After deduplication, 50 unique records remained from an initial 57. Title and abstract screening identified 28 relevant reports; three full texts were unavailable, leaving 25 studies meeting the final eligibility criteria. Protection Motivation Theory (PMT) was central to seven studies, explicit sociotechnical perspectives to four, and the Theory of Planned Behavior (TPB) to three. Cognitive mechanisms were empirically examined in 18 studies, while adversarial factors were contextualized in 21; this evidence reflects model roles rather than effect sizes. Theories combined appraisal and intention with social relationships, organizational factors, technology adoption, and computational modeling. The evidence base ranged from conceptual frameworks to simulated or observed behaviors, with clear distinctions between intentions, self-reported practices, and demonstrated security outcomes. This synthesis offers an evidence-based framework for selecting complementary explanations and guiding future empirical investigations of cybersecurity behavior.
Open AccessHuman Behavior in Cybersecurity from a Sociotechnical Perspective: A Literature Review of Theories and Dimensions
Human behavior in cybersecurity results from the interplay among individual appraisals, organizational conditions, technological factors, and adversarial contexts. This structured literature review investigates the functions of behavioral theories, their extensions and integrations, and the explanatory roles attributed to seven sociotechnical dimensions. We conducted four topic queries in the Web of Science Core Collection in January 2026 and updated the screening in October 2026, targeting journal articles published between 2015 and 2025. After deduplication, 50 unique records remained from an initial 57. Title and abstract screening identified 28 relevant reports; three full texts were unavailable, leaving 25 studies meeting the final eligibility criteria. Protection Motivation Theory (PMT) was central to seven studies, explicit sociotechnical perspectives to four, and the Theory of Planned Behavior (TPB) to three. Cognitive mechanisms were empirically examined in 18 studies, while adversarial factors were contextualized in 21; this evidence reflects model roles rather than effect sizes. Theories combined appraisal and intention with social relationships, organizational factors, technology adoption, and computational modeling. The evidence base ranged from conceptual frameworks to simulated or observed behaviors, with clear distinctions between intentions, self-reported practices, and demonstrated security outcomes. This synthesis offers an evidence-based framework for selecting complementary explanations and guiding future empirical investigations of cybersecurity behavior.
- Open Access
SMA-WSN: Failure-Aware Slime-Mold-Inspired Clustering and Per-Hop Versus End-to-End Delivery Analysis in Agricultural Wireless Sensor Networks
Wireless sensor networks in agricultural fields face abrupt node failures from environmental stress and physical damage, yet most clustering protocols assume fault-free operation. This paper proposes SMA-WSN, a failure-aware clustering protocol inspired by the Slime Mold Algorithm, selecting cluster heads via a tri-objective fitness that combines residual energy, base-station proximity, and local coverage. Evaluated in NS-3.47 against 11 protocols over six agricultural topology archetypes—spatial node-layout patterns derived from Algerian field geometries and densities, not biophysical canopy or soil-moisture propagation models—and five failure rates (18,000 runs, 50 seeds), SMA-WSN achieves the highest per-hop delivery ratio at every failure rate and ranks first overall on that metric, with its largest margins on the heterogeneous, oasis, and sparse-Saharan terrains (T4–T6). We also report an end-to-end delivery ratio, counting a datum only when it reaches the base station: under that stricter definition, all 12 protocols change rank, and SMA-WSN places seventh. We report both and treat the divergence between them as a result in its own right. SMA-WSN achieves the highest per-hop PDR while remaining statistically tied with the best protocols in energy fairness (Jain index 0.922, within one confidence half-width of the leaders GWO-LEACH, PSO-LEACH, and FGO-QL). An ablation over the three fitness terms and their weights shows this advantage is robust to objective weighting and arises from the SMA selection dynamics rather than any single fitness term: no term, including coverage, is individually decisive, so the advantage does not depend on a finely tuned weighting. Disabling the stochastic wrap phase, by contrast, costs 11.4 percentage points of per-hop PDR, identifying it as the mechanism actually responsible for the gain. On First Node Death under failure, SMA-WSN is mid-field rather than dominant; we report this honestly and discuss it as an open question for follow-up work distinguishing energy-exhaustion deaths from abrupt hardware failures.
Open AccessSMA-WSN: Failure-Aware Slime-Mold-Inspired Clustering and Per-Hop Versus End-to-End Delivery Analysis in Agricultural Wireless Sensor Networks
Wireless sensor networks in agricultural fields face abrupt node failures from environmental stress and physical damage, yet most clustering protocols assume fault-free operation. This paper proposes SMA-WSN, a failure-aware clustering protocol inspired by the Slime Mold Algorithm, selecting cluster heads via a tri-objective fitness that combines residual energy, base-station proximity, and local coverage. Evaluated in NS-3.47 against 11 protocols over six agricultural topology archetypes—spatial node-layout patterns derived from Algerian field geometries and densities, not biophysical canopy or soil-moisture propagation models—and five failure rates (18,000 runs, 50 seeds), SMA-WSN achieves the highest per-hop delivery ratio at every failure rate and ranks first overall on that metric, with its largest margins on the heterogeneous, oasis, and sparse-Saharan terrains (T4–T6). We also report an end-to-end delivery ratio, counting a datum only when it reaches the base station: under that stricter definition, all 12 protocols change rank, and SMA-WSN places seventh. We report both and treat the divergence between them as a result in its own right. SMA-WSN achieves the highest per-hop PDR while remaining statistically tied with the best protocols in energy fairness (Jain index 0.922, within one confidence half-width of the leaders GWO-LEACH, PSO-LEACH, and FGO-QL). An ablation over the three fitness terms and their weights shows this advantage is robust to objective weighting and arises from the SMA selection dynamics rather than any single fitness term: no term, including coverage, is individually decisive, so the advantage does not depend on a finely tuned weighting. Disabling the stochastic wrap phase, by contrast, costs 11.4 percentage points of per-hop PDR, identifying it as the mechanism actually responsible for the gain. On First Node Death under failure, SMA-WSN is mid-field rather than dominant; we report this honestly and discuss it as an open question for follow-up work distinguishing energy-exhaustion deaths from abrupt hardware failures.
- Open Access
- 30 June 2026
- Published by Sakarya University Journal of Computer and Information Sciences in Sakarya University Journal of Computer and Information Sciences
- Vol. 9 (3) , 846-867
A Transformer-Assisted Cooperative Caching Framework with Q-Learning Optimization for Internet of Vehicles
The Internet of Vehicles (IoV) is witnessing escalating data traffic that demands sophisticated caching to minimize retrieval latency and sustain Quality of Service. Recent cooperative approaches, notably the Cooperative Caching Strategy with Content Request Prediction (CCCRP) framework grounded in LSTM-based request prediction, struggle to capture the complex, long-range temporal dependencies inherent to dynamic vehicular environments. This paper proposes Transformer-based CCCRP (T-CCCRP), which replaces the LSTM predictor with a Transformer model leveraging self-attention to model long-range dependencies and enable efficient parallelism. The prediction module is integrated with a reinforcement learning controller to optimize cache placement jointly across vehicles and roadside units, thereby aligning predicted popularity with resource constraints. To ensure a realistic and comprehensive evaluation, the proposed framework is assessed under both small-scale and newly introduced large-scale IoV scenarios, involving up to 200 vehicles, multiple caching nodes, and expanded content libraries. The experimental setup further adapts the input sequence length of the prediction model to reflect increased temporal complexity in large-scale environments. Simulation results demonstrate that T-CCCRP consistently outperforms the original CCCRP and conventional caching strategies (LFU and LRU) across all evaluated scales.
- 30 June 2026
- Published by Sakarya University Journal of Computer and Information Sciences in Sakarya University Journal of Computer and Information Sciences
- Vol. 9 (3) , 846-867
Open AccessA Transformer-Assisted Cooperative Caching Framework with Q-Learning Optimization for Internet of Vehicles
The Internet of Vehicles (IoV) is witnessing escalating data traffic that demands sophisticated caching to minimize retrieval latency and sustain Quality of Service. Recent cooperative approaches, notably the Cooperative Caching Strategy with Content Request Prediction (CCCRP) framework grounded in LSTM-based request prediction, struggle to capture the complex, long-range temporal dependencies inherent to dynamic vehicular environments. This paper proposes Transformer-based CCCRP (T-CCCRP), which replaces the LSTM predictor with a Transformer model leveraging self-attention to model long-range dependencies and enable efficient parallelism. The prediction module is integrated with a reinforcement learning controller to optimize cache placement jointly across vehicles and roadside units, thereby aligning predicted popularity with resource constraints. To ensure a realistic and comprehensive evaluation, the proposed framework is assessed under both small-scale and newly introduced large-scale IoV scenarios, involving up to 200 vehicles, multiple caching nodes, and expanded content libraries. The experimental setup further adapts the input sequence length of the prediction model to reflect increased temporal complexity in large-scale environments. Simulation results demonstrate that T-CCCRP consistently outperforms the original CCCRP and conventional caching strategies (LFU and LRU) across all evaluated scales.
- Open Access
A Hybrid Deep Learning Architecture for Content Request Prediction in the Internet of Vehicles
Low-latency content delivery is essential in the Internet of Vehicles (IoV) to support autonomous driving, cooperative perception, and infotainment services. However, rapidly changing vehicular mobility and demand patterns limit the effectiveness of existing content prediction and caching strategies, which often capture either short-term temporal trends or long-range dependencies, but not both. This paper proposes a hybrid deep learning architecture that integrates Long Short-Term Memory (LSTM) networks with Transformer encoders to jointly model fine-grained temporal dynamics and global correlations in content requests. The resulting popularity predictions are incorporated into a reinforcement learning (RL)-based caching policy, enabling proactive and adaptive cache placement at roadside units (RSUs) within an end-to-end optimization framework. Simulation results across representative IoV scenarios show that the proposed approach consistently improves cache hit ratio, retrieval latency, and prediction accuracy compared with LSTM-only, Transformer-only, Least Frequently Used (LFU), and Least Recently Used (LRU) baselines. Ablation studies further demonstrate the complementary strengths of the hybrid components, highlighting improved convergence behavior and robustness under varying demand distributions.
Open AccessA Hybrid Deep Learning Architecture for Content Request Prediction in the Internet of Vehicles
Low-latency content delivery is essential in the Internet of Vehicles (IoV) to support autonomous driving, cooperative perception, and infotainment services. However, rapidly changing vehicular mobility and demand patterns limit the effectiveness of existing content prediction and caching strategies, which often capture either short-term temporal trends or long-range dependencies, but not both. This paper proposes a hybrid deep learning architecture that integrates Long Short-Term Memory (LSTM) networks with Transformer encoders to jointly model fine-grained temporal dynamics and global correlations in content requests. The resulting popularity predictions are incorporated into a reinforcement learning (RL)-based caching policy, enabling proactive and adaptive cache placement at roadside units (RSUs) within an end-to-end optimization framework. Simulation results across representative IoV scenarios show that the proposed approach consistently improves cache hit ratio, retrieval latency, and prediction accuracy compared with LSTM-only, Transformer-only, Least Frequently Used (LFU), and Least Recently Used (LRU) baselines. Ablation studies further demonstrate the complementary strengths of the hybrid components, highlighting improved convergence behavior and robustness under varying demand distributions.
- Open Access
AI-Guided Inflatable Neck Brace for Personalized Cervical Support
Many people suffer from cervical disc herniation, which significantly affects the lives of individuals, causing chronic pain and functional limitations. This paper presents the development and evaluation of an AI-powered inflatable neck collar designed to provide personalized and adaptive support for individuals experiencing neck pain, particularly those with disc herniations. The system seamlessly integrates motion sensors, a robust AI model trained on a dataset of MRI scans, and a custom-designed inflatable collar. The AI model accurately detects disc herniations and segments vertebral structures, enabling real-time, targeted inflation adjustments based on the user’s unique anatomy, posture, and movements. A user-centered design approach ensures a seamless and intuitive user experience, allowing for personalized profile management, control over inflation levels, and data logging for tracking progress. Extensive simulations using 3D models and real-time data flow systems validated the effectiveness of the AI-guided system. Results demonstrated accurate detection and segmentation of disc herniations, robust real-time response, and adaptability to user needs. The proposed system, reviewed and validated by a neurosurgeon, demonstrates significant potential as a novel and effective solution for personalized treatment of neck pain, particularly in cases of disc herniation. Further development and research will focus on expanding the dataset to improve fairness and accuracy for diverse demographics and increasing the robustness and generalizability of the system.
Open AccessAI-Guided Inflatable Neck Brace for Personalized Cervical Support
Many people suffer from cervical disc herniation, which significantly affects the lives of individuals, causing chronic pain and functional limitations. This paper presents the development and evaluation of an AI-powered inflatable neck collar designed to provide personalized and adaptive support for individuals experiencing neck pain, particularly those with disc herniations. The system seamlessly integrates motion sensors, a robust AI model trained on a dataset of MRI scans, and a custom-designed inflatable collar. The AI model accurately detects disc herniations and segments vertebral structures, enabling real-time, targeted inflation adjustments based on the user’s unique anatomy, posture, and movements. A user-centered design approach ensures a seamless and intuitive user experience, allowing for personalized profile management, control over inflation levels, and data logging for tracking progress. Extensive simulations using 3D models and real-time data flow systems validated the effectiveness of the AI-guided system. Results demonstrated accurate detection and segmentation of disc herniations, robust real-time response, and adaptability to user needs. The proposed system, reviewed and validated by a neurosurgeon, demonstrates significant potential as a novel and effective solution for personalized treatment of neck pain, particularly in cases of disc herniation. Further development and research will focus on expanding the dataset to improve fairness and accuracy for diverse demographics and increasing the robustness and generalizability of the system.
- Open Access
A Real-Time Consensus-Free Accident Detection Framework for Internet of Vehicles Using Vision Transformer and EfficientNet
Objectives: Traffic accidents cause severe social and economic impacts, demanding fast and reliable detection to minimize secondary collisions and improve emergency response. However, existing cloud-dependent detection systems often suffer from high latency and limited scalability, motivating the need for an edge-centric and consensus-free accident detection framework in IoV environments. Methods: This study presents a real-time accident detection framework tailored for Internet of Vehicles (IoV) environments. The proposed system forms an integrated IoV architecture combining on-vehicle inference, RSU-based validation, and asynchronous cloud reporting. The system integrates a lightweight ensemble of Vision Transformer (ViT) and EfficientNet models deployed on vehicle nodes to classify video frames. Accident alerts are generated only when both models agree (vehicle-level ensemble consensus), ensuring high precision. These alerts are transmitted to nearby Road Side Units (RSUs), which validate the events and broadcast safety messages without requiring inter-vehicle or inter-RSU consensus. Structured reports are also forwarded asynchronously to the cloud for long-term model retraining and risk analysis. Results: Evaluated on the CarCrash and CADP datasets, the framework achieves an F1-score of 0.96 with average decision latency below 60 ms, corresponding to an overall accuracy of 98.65% and demonstrating measurable improvement over single-model baselines. Conclusions: By combining on-vehicle inference, edge-based validation, and optional cloud integration, the proposed architecture offers both immediate responsiveness and adaptability, contrasting with traditional cloud-dependent approaches.
Open AccessA Real-Time Consensus-Free Accident Detection Framework for Internet of Vehicles Using Vision Transformer and EfficientNet
Objectives: Traffic accidents cause severe social and economic impacts, demanding fast and reliable detection to minimize secondary collisions and improve emergency response. However, existing cloud-dependent detection systems often suffer from high latency and limited scalability, motivating the need for an edge-centric and consensus-free accident detection framework in IoV environments. Methods: This study presents a real-time accident detection framework tailored for Internet of Vehicles (IoV) environments. The proposed system forms an integrated IoV architecture combining on-vehicle inference, RSU-based validation, and asynchronous cloud reporting. The system integrates a lightweight ensemble of Vision Transformer (ViT) and EfficientNet models deployed on vehicle nodes to classify video frames. Accident alerts are generated only when both models agree (vehicle-level ensemble consensus), ensuring high precision. These alerts are transmitted to nearby Road Side Units (RSUs), which validate the events and broadcast safety messages without requiring inter-vehicle or inter-RSU consensus. Structured reports are also forwarded asynchronously to the cloud for long-term model retraining and risk analysis. Results: Evaluated on the CarCrash and CADP datasets, the framework achieves an F1-score of 0.96 with average decision latency below 60 ms, corresponding to an overall accuracy of 98.65% and demonstrating measurable improvement over single-model baselines. Conclusions: By combining on-vehicle inference, edge-based validation, and optional cloud integration, the proposed architecture offers both immediate responsiveness and adaptability, contrasting with traditional cloud-dependent approaches.
- Open Access
- 13 June 2025
- Published by Springer Nature in International Journal of Networked and Distributed Computing
- Vol. 13 (1) , 1-14
Enhancing Green Vehicle Routing Efficiency Through Machine Learning-Optimized Genetic Algorithms
- 13 June 2025
- Published by Springer Nature in International Journal of Networked and Distributed Computing
- Vol. 13 (1) , 1-14
Open AccessEnhancing Green Vehicle Routing Efficiency Through Machine Learning-Optimized Genetic Algorithms
- Open Access
- 31 December 2024
- Published by National Institute of Telecommunications in Journal of Telecommunications and Information Technology
- Vol. 4 (2024) , 43-50
A New Tree Quantum Key Agreement Protocol for Secure Multiparty Communication
The Tree Multiparty Quantum Key Agreement Protocol (TMQKAP) is introduced as a novel solution for secure quantum key agreement among multiple participants, specifically tailored for tree topologies. Based on the BB84 protocol, TMQKAP employs hierarchical tree structures and XOR operations to facilitate efficient and secure key generation. Key elements are exchanged among participants in an equitable manner, ensuring that each participant contributes equally to the generation of the shared key. The protocol demonstrates robust security, effectively defending against both external and internal attacks, and achieves a quantum efficiency of 1/2 (N −1), where N is the number of participants. Thorough security analysis and simulations show TMQKAP’s robustness against various attacks while maintaining high efficiency. Additionally, the protocol is readily implementable with current quantum technologies, utilizing single-photon transmission to facilitate secure key distribution.
- 31 December 2024
- Published by National Institute of Telecommunications in Journal of Telecommunications and Information Technology
- Vol. 4 (2024) , 43-50
Open AccessA New Tree Quantum Key Agreement Protocol for Secure Multiparty Communication
The Tree Multiparty Quantum Key Agreement Protocol (TMQKAP) is introduced as a novel solution for secure quantum key agreement among multiple participants, specifically tailored for tree topologies. Based on the BB84 protocol, TMQKAP employs hierarchical tree structures and XOR operations to facilitate efficient and secure key generation. Key elements are exchanged among participants in an equitable manner, ensuring that each participant contributes equally to the generation of the shared key. The protocol demonstrates robust security, effectively defending against both external and internal attacks, and achieves a quantum efficiency of 1/2 (N −1), where N is the number of participants. Thorough security analysis and simulations show TMQKAP’s robustness against various attacks while maintaining high efficiency. Additionally, the protocol is readily implementable with current quantum technologies, utilizing single-photon transmission to facilitate secure key distribution.
- Open Access
- 30 September 2024
- Published by National Institute of Telecommunications in Journal of Telecommunications and Information Technology
- Vol. 3 (2024) , 65-78
TS-based Mobility-aware Multi-hierarchical Caching Model with Vehicle Clustering and Content Popularity Prediction
This research is concerned with the fusion of artificial intelligence (AI) and machine learning within multi-hierarchical caching systems, specifically targeting vehicular and edge caching domains. This study introduces an innovative architecture harmonizing Thompson sampling learning-based caching policies with advanced vehicle clustering and content-popularity prediction methods (TS-MMCM). Simulations show substantial performance improvements and a big impact of the proposed approach on system efficiency in dynamic network environments. The proposal demonstrates a notable gain in cache hit rates and decreased latency levels, highlighting the potential of AI to improve caching techniques in dynamic network environments.
- 30 September 2024
- Published by National Institute of Telecommunications in Journal of Telecommunications and Information Technology
- Vol. 3 (2024) , 65-78
Open AccessTS-based Mobility-aware Multi-hierarchical Caching Model with Vehicle Clustering and Content Popularity Prediction
This research is concerned with the fusion of artificial intelligence (AI) and machine learning within multi-hierarchical caching systems, specifically targeting vehicular and edge caching domains. This study introduces an innovative architecture harmonizing Thompson sampling learning-based caching policies with advanced vehicle clustering and content-popularity prediction methods (TS-MMCM). Simulations show substantial performance improvements and a big impact of the proposed approach on system efficiency in dynamic network environments. The proposal demonstrates a notable gain in cache hit rates and decreased latency levels, highlighting the potential of AI to improve caching techniques in dynamic network environments.