Journal of Moeslim Research Technik https://www.research.adra.ac.id/index.php/technik <p style="text-align: justify;"><strong>Journal of Moeslim Research Technik</strong> is is a Bimonthly, open-access, peer-reviewed publication that publishes both original research articles and reviews in all fields of Engineering including Civil, Mechanical, Industrial, Electrical, Computer, Chemical, Petroleum, Aerospace, Architectural, etc. It uses an entirely open-access publishing methodology that permits free, open, and universal access to its published information. Scientists are urged to disclose their theoretical and experimental work along with all pertinent methodological information. Submitted papers must be written in English for initial review stage by editors and further review process by minimum two international reviewers.</p> en-US journal@adra.ac.id (Journal of Moeslim Research Technik) journal@adra.ac.id (Admin Journal) Thu, 02 Apr 2026 18:15:54 +0700 OJS 3.2.1.2 http://blogs.law.harvard.edu/tech/rss 60 ALGORITHMIC INTELLIGENCE IN ENGINEERING DESIGN: INTEGRATING MACHINE LEARNING WITH PHYSICAL MODELING https://www.research.adra.ac.id/index.php/technik/article/view/3467 <p>Increasing complexity in engineering systems demands design methodologies that balance computational efficiency, predictive accuracy, and physical reliability. Traditional physics-based simulations ensure mechanistic consistency but are computationally expensive, while purely data-driven machine learning models offer speed yet often lack interpretability and physical compliance. Integrating algorithmic intelligence with physical modeling has therefore emerged as a promising paradigm in advanced engineering design. This study aims to develop and evaluate a hybrid framework that integrates machine learning algorithms with governing physical equations to enhance design performance, robustness, and computational efficiency. A mixed-methods computational design was employed using 15,000 high-fidelity simulation datasets across structural, aerodynamic, and thermal engineering cases. Three modeling configurations—physics-based models, data-driven models, and hybrid physics-informed machine learning models—were comparatively analyzed using performance metrics including mean squared error, R², runtime efficiency, robustness testing, and constraint violation indices. Statistical analyses were conducted to determine significance of performance differences. Hybrid models achieved superior balance, reaching R² = 0.97 with significantly reduced runtime compared to physics-based simulations (p &lt; 0.001), while maintaining substantially lower physical constraint violations than purely data-driven models. Sensitivity and uncertainty analyses confirmed enhanced robustness under parameter perturbation. Algorithmic intelligence integrated with physical modeling represents an epistemologically coherent and practically effective approach, advancing engineering design toward trustworthy, efficient, and physically consistent computational frameworks.</p> Fauzi Erwis, Miku Fujita, I Putu Dody Suarnatha, Amanda Wilson Copyright (c) 2026 Fauzi Erwis, Miku Fujita, I Putu Dody Suarnatha, Amanda Wilson https://creativecommons.org/licenses/by-sa/4.0 https://www.research.adra.ac.id/index.php/technik/article/view/3467 Fri, 03 Apr 2026 00:00:00 +0700