Parametric Modeling for Intelligent, Automated Design of Footing-Type Surface Foundations: Optimized Predesign

The design of isolated shallow footings is commonly performed through iterative procedures based on the designer’s experience, often resulting in conservative solutions. This paper presents a computational methodology that integrates structural, geotechnical, and geometric criteria for automated and optimized design of isolated footings. The formulation minimizes concrete volume subject to constraints based on Brazilian standards NBR 6118 and NBR 6122. The problem is solved using Efficient Global Optimization (EGO), combining Gaussian Process Regression, Expected Improvement, and a Genetic Algorithm as the internal optimizer. The methodology was implemented in the FundaAI web platform and applied to examples with one, two, and three footings. The results indicated volume reductions compared with Monte Carlo sampling and highlighted limitations in configurations with severe footing overlap.