An Improved Two-Stage Bottom-Up Optimization Approach for Pavement Maintenance and Rehabilitation Decision Making

Abstract

Existing two-stage bottom-up optimization (TSBUO) approaches have a simple combination of optimization paths at the first stage, leading to the possibility of missing the optimal plan at the second stage with complex network level constraints. To address this drawback, this study proposes an improved TSBUO approach for pavement maintenance and rehabilitation (M&R) decision-making for multi-year planning periods. The two stages are the segment and network levels, with the optimization objective of maximizing the effectiveness-cost ratio (E/C). The optimal and suboptimal M&R paths for each road segment under different initial M&R actions are searched at the segment level using the backtracking algorithm, and the composed optimal M&R plan is developed at the network level using a genetic algorithm (GA). The results are obtained from a case study. Compared with the actual engineering plan, the proposed approach results in an M&R plan with 5.9% higher average pavement performance at a lower cost. Meanwhile, compared with two existing TSBUO approaches for pavement M&R decision-making, the proposed approach results in a greater E/C, demonstrating the better optimization-seeking ability of the proposed approach. Therefore, this approach can help decision makers to develop better M&R plans.