Model formulation and comparison for network-level pavement maintenance strategies

This study aims to quantitatively evaluate different maintenance strategies for transportation infrastructure such as pavement, bridges, and railway. Using highway pavements as an example and the pavement performance model in the literature that considers deterioration trends responding to traffic and pavement age and maintenance actions with heterogeneous effects as basis, the study develops mixed-integer linear programming models for various maintenance strategies commonly adopted in practice. The strategies include scheduled, worst-first, best-first, and threshold-based. In order to represent the practical situation where engineering judgment is incorporated in the worst-first maintenance, a hybrid strategy is further developed. In order to efficiently solve the models for different strategies, the study proposes a three-stage solution approach. Finally, the case study of a hypothetical network is conducted. Different maintenance strategies are applied given different budget levels, traffic loadings, and initial pavement conditions. The results of the case study are reasonable, which shows that the models have potential to help the maintenance agencies to evaluate maintenance strategies and to improve performance and reduce budget for transportation infrastructure.