Introducing an Optimized Intelligent Model for Evaluation of Bearing Capacity of Shallow Foundations Over Granular Soil

The term "bearing capacity" refers to the maximum stress which can be applied to the soil through a foundation without any shear failure. Owing to complexities of soil behavior, precise calculation of this stress using analytical methods is not possible and the available approaches only give an approximation of bearing capacity. Intelligent methods have attracted a great attention as powerful tools in solving engineering problems particularly in the fields where conventional approaches are not able to give accurate answers. Recently, a number of intelligent methods have been applied to estimate the bearing capacity of shallow foundations. However, the models developed in previous studies are not optimized properly and they are designed mainly by trial and error. In current study, optimization of intelligent models for evaluation of the bearing capacity has been investigated. In this regard, the previous methods proposed by other researchers are assessed first. Subsequently, it is aimed to obtain optimized complexity of the models in order to prevent over-estimating or under-estimating the bearing capacity of shallow foundations over granular soil in the cases which are not considered during training phase.