arXiv subjects

Amel Ourici

Publications and source records attributed to Amel Ourici.

5 recordsLinked to original sources

arXiv2026

NaviGNN: Multi-Agent Reinforcement Learning and Graph Neural Network for Sustainable Mobility in Futuristic Smart Cities

This paper investigates the feasibility of human mobility in extreme urban morphologies characterized by high-density vertical structures and linear city layouts. To assess whether agents can navigate efficiently within such unprecedented topologies, we develop a hybrid simulation framework integrating agent-based modeling, reinforcement learning (RL), supervised learning, and graph neural networks (GNNs). The simulation captures multi-modal transportation behaviors across multiple vertical levels and varying density scenarios, using both synthetic data and real-world traces from high-density cities. Experimental results show that the fully integrated AI architecture enables agents to achieve an average commute time of 7.8-8.4 minutes, a satisfaction rate exceeding 89\%, and a reachability index above 91\%, even during peak congestion periods. Ablation studies indicate that removing intelligent modules such as RL or GNNs significantly degrades performance, with commute times increasing by up to 85\% and reachability dropping below 70\%. Baseline comparisons against Dijkstra, A*, DQN, and standard GCN further confirm the superiority of the proposed model across all mobility and sustainability metrics. Environmental modeling demonstrates low energy consumption and minimal CO2 emissions when electric transportation modes are prioritized. These findings suggest that efficient and sustainable mobility in extreme urban environments is achievable, provided that adaptive AI systems, intelligent infrastructure, and real-time feedback mechanisms are effectively implemented.

cs.AI↗

arXiv2026

Green Energy Management for Sustainable Data Centers Using Deep Reinforcement Learning

The exponential growth of digital services has positioned data centers among the most energy-intensive infrastructures in the modern economy, raising critical concerns regarding operational costs, carbon emissions, and the sustainable integration of renewable energy sources. This paper proposes a novel Deep Reinforcement Learning (DRL)-based energy management framework for data centers, designed to dynamically coordinate solar photovoltaic generation, wind power, battery storage systems, and conventional grid electricity under highly stochastic operational conditions. The proposed framework formulates the energy management problem as a Markov Decision Process and employs a Proximal Policy Optimization (PPO) agent augmented with a hybrid Long Short-Term Memory and temporal attention architecture, enabling accurate modeling of workload dynamics and renewable generation variability. A multi-objective reward function jointly minimizes energy costs, carbon emissions, and service-level agreement (SLA) violations while promoting efficient storage utilization. Extensive experiments conducted on three datasets demonstrate that the proposed framework achieves a 38\% reduction in energy costs compared to rule-based heuristics and outperforms the strongest DRL baseline by 4.6\%, while maintaining an SLA violation rate as low as 1.5\% and an energy efficiency of 83.7\%. Ablation studies confirm the individual contribution of each architectural component, and hyperparameter sensitivity analysis validates the robustness of the approach across a range of configurations.

cs.LG↗

arXiv2026

MOSAIC: Multi-Domain Orthogonal Session Adaptive Intent Capture for Prescient Recommendations

Capturing user intent across heterogeneous behavioral domains stands as a fundamental challenge in session-based recommender systems. Yet, existing multi-domain approaches frequently fail to isolate the distinct contribution of cross-domain interactions from those arising within individual domains, limiting their ability to build rich and transferable user representations. In this work, we propose MOSAIC, a Multi-Domain Orthogonal Session Adaptive Intent Capture framework that explicitly factorizes user preferences into three orthogonal components: domain-specific, domain-common, and cross-sequence-exclusive representations. Our approach employs a triple-encoder architecture, where each encoder is dedicated to one preference type, enforced through domain masking objectives and adversarial training via a gradient reversal layer. Representational alignment and mutual independence constraints are jointly optimized to ensure clean preference separation. Additionally, a dynamic gating mechanism modulates the relative contribution of each component at every timestep, yielding a unified and temporally adaptive session-level user representation. We conduct extensive experiments on two large-scale real-world benchmarks spanning multiple domains and interaction types. The ablation study validates that each component domain-specific encoding, domain-common modeling, cross-sequence representation, and dynamic gating contributes meaningfully to the overall performance. Experimental results demonstrate that MOSAIC consistently outperforms state-of-the-art baselines in recommendation accuracy, while simultaneously providing interpretable insights into the interplay between domain-specific and cross-domain preference signals. These findings highlight the potential of orthogonal preference decomposition as a principled strategy for next-generation multi-domain recommender systems.

cs.IR↗

arXiv2026

FreeGNN: Continual Source-Free Graph Neural Network Adaptation for Renewable Energy Forecasting

Accurate forecasting of renewable energy generation is essential for efficient grid management and sustainable power planning. However, traditional supervised models often require access to labeled data from the target site, which may be unavailable due to privacy, cost, or logistical constraints. In this work, we propose FreeGNN, a Continual Source-Free Graph Domain Adaptation framework that enables adaptive forecasting on unseen renewable energy sites without requiring source data or target labels. Our approach integrates a spatio-temporal Graph Neural Network (GNN) backbone with a teacher--student strategy, a memory replay mechanism to mitigate catastrophic forgetting, graph-based regularization to preserve spatial correlations, and a drift-aware weighting scheme to dynamically adjust adaptation strength during streaming updates. This combination allows the model to continuously adapt to non-stationary environmental conditions while maintaining robustness and stability. We conduct extensive experiments on three real-world datasets: GEFCom2012, Solar PV, and Wind SCADA, encompassing multiple sites, temporal resolutions, and meteorological features. The ablation study confirms that each component memory, graph regularization, drift-aware adaptation, and teacher--student strategy contributes significantly to overall performance. The experiments show that FreeGNN achieves an MAE of 5.237 and an RMSE of 7.123 on the GEFCom dataset, an MAE of 1.107 and an RMSE of 1.512 on the Solar PV dataset, and an MAE of 0.382 and an RMSE of 0.523 on the Wind SCADA dataset. These results demonstrate its ability to achieve accurate and robust forecasts in a source-free, continual learning setting, highlighting its potential for real-world deployment in adaptive renewable energy systems. For reproducibility, implementation details are available at: https://github.com/AraoufBh/FreeGNN.

cs.LG↗

arXiv2025

From Electrochemical Energy Storage to Next-Generation Intelligent Battery Technologies for Electric Vehicles: A Survey

This study provides a comprehensive overview of recent advances in electrochemical energy storage, including Na+ -ion, metal-ion, and metal-air batteries, alongside innovations in electrode engineering, electrolytes, and solid-electrolyte interphase control. It also explores the integration of machine learning, digital twins, large language models and predictive analytics to enable intelligent battery management systems, enhancing performance, safety, and operational longevity. Key challenges, research gaps, and future prospects are addressed, highlighting opportunities presented by hybrid chemistry, scalable manufacturing, sustainability, and AI-driven optimization. This survey aims to provide researchers, engineers, and industry profesionnals with a comprehensive understanding of next-generation battery technologies for the evolving electric vehicles sector.

eess.SY↗