Prediction of Ultimate Bearing Capacity of Shallow Foundations on Cohesionless Soils: An Evolutionary Approach

摘要

This study proposes an innovative mathematical formula that uses multigene Genetic Programming (GP), a recently developed soft computing technique, to predict the ultimate bearing capacity of shallow foundations on cohesionless soils. The real performance of previously developed approaches is also investigated. The multigene GP-based formula was calibrated and validated using an experimental database consisting of approximately one hundred load tests. One half of the data was obtained from full-scale foundations and the other half was obtained from small-scale laboratory footing load tests. The results revealed that the proposed formula by multigene GP could predict the ultimate bearing capacity precisely under the described conditions with a coefficient of correlation of about 98%. Additionally, a comprehensive parametric study on the proposed multigene GP-based formula was conducted to confirm the new methodology's geotechnical aspects.