Neurofuzzy networks applied to settlement of shallow foundations on granular soils

Abstract

In recent years, artificial neural networks (ANNs) have been applied successfully to many aspects of geotechnical engineering and its related fields. In the majority of these applications, multi-layer perceptrons (MLPs) that are trained with the back-propagation algorithm are used. This can be attributed to the fact that MLPs trained with back-propagation have a high capability of data mapping. However, one shortcoming of this type of ANN is that knowledge acquired during training is stored in the connection weights in a complex manner that is often difficult to interpret. Consequently, the rules governing the relationships between the network input/output variables are difficult to quantify, and thus ANNs are often criticized for being black boxes. One way to overcome this problem is to use neurofuzzy networks. Neurofuzzy networks combine the explicit knowledge representation of fuzzy systems with the potential power of ANN learning algorithms. Neurofuzzy networks can be trained by processing data samples to perform input/output mappings, similar to the way MLPs do, with the additional benefit of being able to translate the acquired knowledge into a set of transparent fuzzy rules that clearly describe the model input/output relationships. A review of the literature indicates that neurofuzzy networks are new tools in the field of geotechnical engineering. In this paper, the feasibility of using neurofuzzy networks in the field of geotechnical engineering is investigated for a case study of predicting settlement of shallow foundations on granular soils. The type of neurofuzzy network that is used in this work is a B-spline network, which uses the adaptive spline modeling of observation data (ASMOD) algorithm to automatically optimize the model structure and the number of model inputs. The results indicate that neurofuzzy networks are a useful technique that are: (i) capable of accurately predicting the settlement of shallow foundations on granular soils, and (ii) able to provide a transparent understanding of the relationship between settlement and the factors affecting it.