Risk score inference for bridge maintenance projects using genetic fuzzy weighted pyramid operation tree

In bridge maintenance, risk assessment is critical to prioritizing project work to minimize related risks and costs. However, the conventional method of risk assessment relies heavily on subjective judgments. The Genetic Fuzzy Weighted Pyramid Operation Tree (GFWPOT) was developed in this study to build a formula to solve the uncertainty problem and provide accurate prediction to bridge maintenance risk assessment. GFWPOT applies a pyramid-shaped, 4-connected operation tree (OT) that integrates a genetic algorithm (GA) and fuzzy theory. The performance of GFWPOT was compared against that of other AI models using the reference index (RI) metric, which summarizes three different performance measures (RMSE, MAE, and MAPE). GFWPOT achieved the highest evaluation criteria, RI = 0.93 and RI = 0.95, respectively, for the training and testing datasets. Based on its excellent predictive performance, GFWPOT is recommended as a viable tool to assist decision-makers to develop and execute effective bridge maintenance plans.