Integration of Support Vector Machine with Particle Swarm Optimization in Quality Modeling of CFG Composite Foundation

The purpose of this paper is to develop models to predict the quality (bearing capacity of foundation and pile completeness) of the Cement Fly-ash Gravel (CFG) composite foundation. To ensure the models are useful to both designer and construction of composite foundation, the study employs a broad range of bearing capacity of foundation and pile completeness variables through investigating the load transfer mechanism and reinforcement mechanism of mattress layer mechanism, the pile, layer, replacement ratio, construction machinery and other aspects of CFG pile composite foundation systematic and comprehensively. In addition, the research develops a model integrating of support vector machine with particle swarm optimization (PSO-SVM) for predicting the bearing capacity and pile completeness of the Cement Fly-ash Gravel foundation. This paper contributes to the scarce literature on design and construction of CFG pile composite foundation. The models developed in this paper are useful to give those design and construction of CFG pile composite foundation an early warning. And through specific examples of analysis, the model practical application effect was test. It was concluded that using the established model for CFG pile composite foundation quality verification, is of feasibility and validity.