Computer Engineering, Artificial Intelligence, Computational Theory and Mathematics, Management Science and Operations Research
44
Scopus Publications
Scopus Publications
Infection Aware Hyper-Heuristic Framework for Hospital Room–Patient Matching Kassem Danach, Wael Hosny Fouad Aly, Chadi Fouad Riman Algorithms, 2026 The assignment of hospital rooms to patients is a critical operational decision that has a direct impact on patient safety, infection control, and staff workload. This study introduces HRPM–IRC, an epidemiology-aware hyper-heuristic framework developed to optimize room–patient matching by minimizing the risk of nosocomial infections, reducing travel and specialty mismatch costs, and promoting equitable nurse workload distribution. A mixed-integer linear programming model is formulated to capture infection transmission probabilities, isolation and cohorting requirements, and multi-ward capacity constraints. On top of this model, a bio-inspired hyper-heuristic adaptively selects and refines low-level heuristics, including cohort-first greedy allocation, risk-gradient swaps, and pathogen-aware local MILP refinement, on the basis of contextual epidemiological indicators and reinforcement learning. The framework was validated using a real-world dataset obtained from a tertiary hospital in Lebanon, comprising 142 anonymized patient admissions, 35 rooms, and six nursing teams. Results demonstrate that HRPM–IRC consistently reduces modeled infection risk and workload imbalance by up to forty percent compared to conventional assignment heuristics while maintaining near-real-time decision-making capabilities suitable for dynamic hospital operations. These findings underscore the effectiveness of epidemiology-aware hyper-heuristics in enhancing hospital resilience, improving infection prevention, and supporting fair resource utilization in data-limited healthcare environments typical of Lebanon and other middle-income countries.
An explainable data-driven optimization framework for industrial predictive maintenance scheduling Kassem Danach, Hassan Harb, Hamza Issa, Louai Saker Results in Engineering, 2026 Unexpected equipment failures in industrial systems can lead to significant production downtime, increased operational costs, and reduced asset life cycles. Leveraging recent advancements in data science and intelligent decision-support systems, this paper proposes an explainable data-driven optimization and mathematical modeling framework for predictive maintenance scheduling in industrial environments. The framework integrates sensor-based condition monitoring data with machine learning models to forecast potential equipment failures before they occur. A supervised learning approach, implemented using gradient boosting and temporal feature engineering, predicts the remaining useful life (RUL) and failure probability of critical assets. An explainability layer, based on SHAP (SHapley Additive exPlanations) values, provides interpretable insights into the most influential factors contributing to predicted failures, enabling maintenance engineers to validate model outputs and trust automated recommendations. The predictive outputs are then embedded into a mixed-integer linear programming (MILP) model to generate optimal maintenance schedules that minimize total downtime and maintenance costs while satisfying operational constraints such as resource availability and production deadlines. The proposed framework is validated using a combination of publicly available predictive maintenance datasets and real-world industrial sensor data. Experimental results demonstrate a reduction of up to 22% in unplanned downtime and 15% in maintenance costs compared to reactive and preventive maintenance strategies, while maintaining high interpretability for domain experts. This integrated approach highlights the potential of combining explainable AI, predictive analytics, and mathematical optimization for sustainable and efficient industrial operations.
AI-powered hub optimization: a reinforcement learning and graph-based approach to scalable blockchain networks Kassem Danach, Hassan Rkein, Alaaeddine Ramadan, Hassan Harb, Bassam Hamdar Iaes International Journal of Artificial Intelligence, 2026 Blockchain networks face persistent scalability challenges, including network congestion, high latency, and transaction costs. To address these limitations, this study proposes an AI-driven hub location optimization framework that integrates reinforcement learning (RL), mixed integer linear programming (MILP), and graph neural networks (GNNs). The RL-based hub selection dynamically identifies optimal supernode placement, while MILP ensures cost-efficient transaction routing, and GNNs predict flow patterns for proactive congestion management. Experimental results on Ethereum and Bitcoin datasets demonstrate significant improvements, including a 58.6% reduction in transaction latency, 28.9% gas fee savings, and 41.5% congestion reduction. Beyond performance gains, statistical tests confirm the significance of these improvements, and ablation studies highlight the complementary role of each component.
Digital Dewaxing of Raman Hyperspectral Images: Application to Paraffin-Embedded Skin Biopsies Abbas Rammal, Rabih Assaf, Kassem Danach Journal of Raman Spectroscopy, 2026 Raman spectral imaging is a powerful tool for studying the molecular composition of biological samples, typically preserved in paraffin to protect their molecular structures. However, the intense signal from paraffin can interfere with the analysis, necessitating the removal of pure paraffin pixels for accurate imaging. Extended multiplicative signal correction (EMSC) has recently emerged as an effective method to neutralize the paraffin signal in recorded Raman spectral images. This article introduces a novel methodology combining EMSC with multivariate analysis techniques to separate paraffin and tissue pixels. Blind source separation techniques such as independent component analysis (ICA), non‐negative matrix factorization (NMF), principal component analysis (PCA), and singular value decomposition (SVD) are applied to Raman images acquired solely on paraffin to model and extract the pure paraffin component accurately. To validate our approach, we employ k‐means clustering on the corrected spectra obtained through the proposed method and compare the results with those from the traditional EMSC method that does not utilize pure paraffin components modeled by blind source separation techniques. The deparaffinized spectra are then used to construct Raman images of human tissues, which are compared with hematoxylin‐eosin (H&E) stained tissues for verification. This study demonstrates the potential of Raman spectroscopy, combined with EMSC and blind source separation techniques, as a digital dewaxing tool for analyzing paraffin‐embedded tissues.
Quantum-Inspired Hub Location Framework for Drug-Target Interaction Networks Kassem Danach, Jomana Al Haj Hassan, Ali Raad Proceedings 5th International Conference on Informatics and Software Engineering Iisec 2026, 2026 This paper introduces a novel mathematical model and a quantum-inspired hyperheuristic (QIHH) framework for national-scale hub location optimization. Motivated by the structural analogy between postal logistics and drug-target interaction networks, the proposed model reformulates the classical HLP as a layered network with interaction-based flow dynamics. A new cost function integrates affinity-weighted connectivity and modularity penalties, reflecting DTI-like robustness. The QIHH dynamically selects and tunes low-level search heuristics (annealing, evolutionary, and constructive operators) through an adaptive quantum-state controller, enabling efficient exploration of the discrete solution space. Using the Australian Post dataset, our results show superior cost and scalability compared to classical MILP and metaheuristic benchmarks. The framework reveals emergent modular hub topologies that parallel DTI network properties, bridging complex biological and logistical optimization paradigms.
Carbon-Aware Scheduling in Cloud Computing Operations: A Multi-Objective Optimisation Approach Kassem Danach, Kassem Hamze, Hassan Harb, Hassan Kanj Iet Smart Grid, 2026 The rapid expansion of cloud computing has intensified the environmental impact of large‐scale data centres, which now represent a significant portion of global electricity consumption. Traditional scheduling strategies typically optimise performance or cost, disregarding the fluctuating carbon intensity of regional power grids. This study proposes a dynamic carbon‐aware scheduling framework that integrates real‐time carbon intensity forecasting with multi‐objective optimisation and adaptive rolling‐horizon control. The proposed model simultaneously minimises operational cost and greenhouse gas emissions by intelligently shifting computational workloads across time and geography in response to renewable energy availability. The framework combines an ensemble forecasting module, using long short‐term memory (LSTM) and gradient boosting regression, with a mixed‐integer linear programming (MILP) model solved via the ‐constraint method. It adaptively updates scheduling decisions based on updated carbon forecasts and workload arrivals. Experimental validation on real datasets from the UK National Grid and Google Cloud workload traces demonstrates an average reduction in emissions, a improvement in cost efficiency and less than performance degradation compared to conventional schedulers. Pareto front analysis further reveals actionable trade‐offs between economic efficiency and environmental sustainability. The results confirm that integrating operational research with carbon intelligence enables cloud infrastructures to become both cost‐effective and climate‐aligned.
Adaptive Hyper-Heuristics for Smart Logistics Optimization Kassem Danach, Hasan Fayyad-Kazan, Wissam Khalil, Samir Haddad, Jinane Sayah Lecture Notes in Electrical Engineering, 2026 The complexity of logistics combinatorial optimization problems including vehicle routing, warehouse scheduling, and dynamic delivery has increased because of rising demand and evolving constraints. Metaheuristics show effectiveness but need problem-specific tuning and demonstrate limited general applicability. This research presents a learning-based hyper-heuristic framework which operates at high abstraction levels to select or generate low-level heuristics through dynamic decision-making based on problem features and real-time performance feedback. The proposed system uses reinforcement learning to select heuristics while pursuing adaptability, scalability and domain independence. Additionally, the framework demonstrates its effectiveness through benchmark dataset experiments, which show better solution quality and improved computational efficiency and robustness compared to traditional metaheuristics. Moreover, the framework shows its capability to perform automated decision-making while minimizing human involvement and demonstrating effective adaptation to changing logistics environments. Finally, this research presents an adaptable intelligent optimization system which enhances operational efficiency and resilience in smart supply chain systems.
Energy-Aware Hub Location via Reinforcement Learning and Thermodynamic Optimization Kassem Danach, Samir Haddad, Wael Hosny Fouad Aly, Chadi Kallab, Jinane Sayah IEEE Open Journal of Intelligent Transportation Systems, 2026 The Hub Location Problem (HLP) remains a cornerstone of strategic decision-making in logistics and transportation networks. However, emerging challenges driven by climate variability, energy consumption constraints, and carbon emission regulations are reshaping the complexity of these classical models. This paper introduces a novel Energy-Aware Dynamic Hub Location Problem (EA-DHLP) that integrates climate-driven demand variations, renewable energy availability, and carbon penalties into the hub network design. We formulate the core EA-DHLP as a mixed-integer <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">linear</i> program (MILP) by treating climate and energy series as exogenous parameters and isolating any nonlinear couplings as optional MINLP extensions. To solve large instances of this MILP-based and dynamic problem, we propose a Hybrid Reinforcement Learning and Thermodynamic Hyper-Heuristic (RL-THH) framework. The reinforcement learning agent adaptively selects from a pool of low-level heuristics based on real-time performance feedback, while a thermodynamic-inspired temperature control mechanism dynamically balances exploration and exploitation throughout the search process. We benchmark RL-THH against (i) exact MILP solving on small/medium instances and (ii) advanced metaheuristics/hyperheuristics (including Artificial Bee Colony, Grey Wolf Optimizer, and an RL-only HH variant) under matched computational budgets, with nonparametric significance testing (Wilcoxon with Holm–Bonferroni correction) and effect-size reporting (Cliff’s δ). Extensive computational experiments are conducted on both synthetic and real-world datasets, including the Australian Post and CAB datasets, incorporating climate and energy data. Results show statistically significant improvements in minimizing total costs, reducing energy consumption, and lowering carbon emissions when compared to traditional metaheuristics and the aforementioned state-of-the-art baselines. For small/medium instances, the exact MILP provides optimality certificates that contextualize RL-THH’s performance; for larger instances, RL-THH scales while preserving solution quality and stability. This study offers a scalable, adaptive, and environmentally conscious solution framework for modern hub location problems under sustainability and uncertainty considerations.
Hyper-heuristic driven smart contracts for DeFi: a framework for dynamic rule optimization and adaptive executions Kassem Danach, Hassan Rkein, Ahmad Farroukh, Ziad E. L. Balaa, Samir Haddad Frontiers in Blockchain, 2026 The static and hard-coded logic of smart contracts in Decentralized Finance (DeFi) platforms significantly limits their adaptability in dynamic and volatile market environments. To address this challenge, we propose a novel hyper-heuristic driven framework that enables real-time rule optimization within smart contracts, thereby enhancing responsiveness, gas efficiency, and operational robustness. The framework features a two-layer architecture: a reinforcement learning-based high-level controller selects appropriate low-level rule heuristics from a domain-specific library based on evolving transaction contexts and on-chain data. Implemented and evaluated on Uniswap v2 and Aave v3 protocols, the system dynamically optimizes parameters such as slippage tolerance, gas usage thresholds, and loan-to-value ratios. Experimental results on real-world datasets show significant performance improvements, including a 45.6% increase in transaction success rate, 28.3% reduction in average gas consumption, and 38.4% drop in liquidation events under market stress scenarios. This research demonstrates the feasibility and advantages of embedding intelligent, adaptive decision-making mechanisms within DeFi smart contracts, opening new pathways toward autonomous, resilient, and regulation-aligned blockchain systems.
A hyper-heuristic optimization and communication framework for sustainable agricultural decision-making: A case study on Lebanese farms Kassem Danach, Hassan Harb, Hamza Issa Journal of Agriculture and Food Research, 2025 Agricultural systems in Lebanon face intertwined constraints, water scarcity, fragmented landholdings, first-mile logistics frictions, and climatic variability, that limit the effectiveness of single-technique optimizers. We propose a selection-based hyper-heuristic decision-support framework that learns to choose among diverse low-level heuristics via a multi-armed bandit policy (softmax exploration–exploitation) with a simulated-annealing-inspired move acceptance. Using a real-world dataset from 15 smallholder farms in the Beqaa Valley across three seasons (120 decision variables; 50 constraints), we jointly address crop planning, irrigation scheduling, and first-mile routing formulated as capacitated vehicle routing with time windows (CVRPTW). Empirically, the framework outperforms Genetic Algorithm, Simulated Annealing, and Tabu Search with 9.7–15.8% higher normalized objective values across problems, trading a modest 18.5% average runtime increase versus the fastest baseline (SA). The gains materialize through concrete mechanisms that directly affect decisions: (i) higher water-use efficiency by prioritizing parcels with the largest marginal yield response under scarcity; (ii) resilient crop mixes that balance water-intensive and drought-tolerant varieties across fragmented parcels; and (iii) clustered farm pickups that reduce first-mile distance and distribution cost. These outputs translate into actionable seasonal crop-mix plans, weekly irrigation calendars, and daily pickup routes. The learned policies exhibit interpretable patterns aligned with agronomic intuition, improving trust and adoption potential. While developed in Lebanon, the framework is directly transferable to smallholder, water-stressed contexts with heterogeneous parcels and volatile logistics, offering an adaptive, interpretable, and scalable pathway toward more sustainable resource allocation and yield-oriented decisions. • Traditional optimization methods often struggle to adapt to dynamic agricultural systems. • A novel hyper-heuristic optimization framework for agricultural decision-making is proposed. • Our framework develops a scalable solution to optimize key agricultural operations. • The framework tackles problems such as crop planning, irrigation optimization, and distribution logistics under uncertainty.
"Smart"Integration of Statistical Approaches in AI Development Chadi Kallab, Samir Haddad, Rushdi Zaiter, Nisrine Turkey, Imad Zakhem, Jinane Sayah, Kassem Danach, Joseph Merhej International Conference on Computer and Applications Icca 2025 Proceedings, 2025
