Optimizing bridge decks maintenance strategies based on probabilistic performance prediction using genetic algorithm

Bridges are important structures in transportation networks, and their maintenance is essential to public safety. Therefore, there is a critical need for research about evaluating the condition of existing bridges, investigating rehabilitation methods and organizing a management model for these bridges. Bridge Management Systems offer an effective decision-making tool for prioritizing maintenance, repair and rehabilitation (MR&R) activities taking into consideration such factors as budget constraints, suitability of MR&R methods, type and severity of bridge damages, safety, and user cost. In this research a multi-objective Genetic Algorithm is proposed to find the optimal long-term MR&R strategies for a set of reinforced concrete bridge decks based on the current status of the bridges, the applicability of several MR&R methods and their recovering effects, safety of the network, and the available budget. In this process, uncertainties associated with performance and safety have been modeled. The proposed methodology is demonstrated using a case study about bridges in Montreal partially based on real data obtained from the Ministry of Transportation of Quebec