Neural networks for evaluating construction technology

Abstract

This chapter first briefly reviews some neural network applications developed in the area of construction engineering and management before focusing on the topic of evaluating construction technology. Alternative methods and equipment can be used for most construction operations considered on building projects. The decision to adopt a new technology and implement it in place of existing technical solutions should be based on its relative merits while considering the priorities and needs of the users. This chapter presents an artificial neural network-based approach to evaluating innovative construction equipment. The approach is intended to help decision-makers in contracting firms and technology developers to implement new technology and improve operation performance. Two neural network applications are proposed to handle two aspects of construction equipment technology: production capacity and overall acceptability. One reason for using neural networks as an evaluation mechanism for new construction technologies is the networks' ability to learn to perform complex input-output mapping. In this chapter, procedures for model-structuring, data collection, and network training and testing are provided, and illustrated using technology evaluation examples. Since sufficient accuracy is achieved in these examples with limited data collection effort, it is concluded that neural networks have a potential for becoming an efficient means of evaluating new construction technologies.