Comparing the use of artificial neural networks and case-based reasoning in modeling bridge deterioration

Bridge Management Systems (BMSs) have been developed to optimize maintenance, rehabilitation, and replacement decisions for bridge networks under budget constraints. The success of these systems depends greatly on their ability to predict the future condition of bridges/bridge components in an accurate and timely fashion. Current BMSs employ deterioration models, such as regression models and Markovian models, for that purpose. Since the late 1990's, artificial intelligence approaches, such as Artificial Neural Networks (ANNs) and Case-Based Reasoning (CBR) have been proposed to develop deterioration models that eliminate the limitations of existing models based on their ability to learn from data and to model complex relationships. In this paper, a comparison of using ANNs and CBR in modeling bridge deterioration is carried out using bridge deck data obtained from the Ministry of Transportation of Quebec. The objective of this comparison is to determine the pros and cons of the two approaches and to guide transportation agencies in selecting the approach that best suits their needs.