Reliability-Based Design Optimization of Shallow Foundation on Cohesionless Soil Based on Surrogate-Based Numerical Modeling

Abstract This paper develops an approach that utilizes kriging surrogate models for the reliability-based design of geotechnical structures. Surrogate models improve the computational efficiency and reduce the complexity of implementation to model the deterministic numerical models. This procedure involves the calibration of a kriging model, using samples generated by any deterministic geotechnical numerical model, which can be used in geotechnical design and optimization. Reliability-based design and optimization using a numerical model involve multiple calls of the numerical model, which is made possible with the kriging surrogate model. The procedure is illustrated through a simple shallow foundation example. The results demonstrate that reliability-based design and optimization based on surrogate modeling are efficient tools that can be used to design geotechnical structures.