Settlement-based cost optimization of geogrid-reinforced pile-supported foundation

Abstract

Cost optimization of Geogrid-Reinforced Pile-Supported Foundation (GRPSF) requires the minimum construction cost among all design alternatives within both ultimate limit state (ULS) and serviceability limit state (SLS) criteria. Usually, the optimization is conducted by selecting a limited number of design alternatives based on experience and then comparing them, which often does not lead to the real optimal design. This paper presents a novel optimization framework to systematically determine the design parameters to achieve the minimum construction cost for GRPSF, considering both ULS and SLS constraints that are relevant to post-construction performance and constructability. This framework is a hybrid of surrogate modeling and Finite Element Method (FEM) to calculate the post-construction settlement of GRPSF and search for the optimal design. Genetic Algorithm improved Black Hole Algorithm (BH-GA) was developed to determine the optimal values of design variables, including pile length and spacing, pile cap geometry, and geogrid layers and layout. The proposed approach can quickly identify the optimal design by exhausting all possible combinations of design parameters. Two well-documented case histories of GRPSF were redesigned using this framework, which validated its applicability and effectiveness in optimizing the design of GRPSF.