Trenchless Technologies Decision Support System Using Integrated Hierarchical Artificial Neural Networks and Genetic Algorithms

Abstract

The tunneling industry involves countless number of variables and complexities that have to be considered when selecting the construction method to be applied in different types of projects. In addition, the availability of different trenchless technologies makes it difficult to select the most suitable trenchless technology to be used. This paper introduces a framework for developing a Trenchless Technology Decision Support System (TTDSS) using a newly-introduced technique "Hierarchical Artificial Neural Networks (ANN)". The system integrates the concept of hierarchies with the ANN, taking into consideration the direct effect of the factors on each hierarchical selection. Sixty projects were introduced to the HANN, 80% of them were used as training cases and the remaining 20% were used for testing. Results indicated the potential of TTDSS in supporting trenchless technology specialists in their selection decisions, where the error percentage did not exceed 5%.