New approach to programming maintenance activities for concrete bridge decks

This paper presents an approach to programming maintenance alternatives for a network of concrete bridge decks using genetic algorithms and Markovian performance prediction models. Genetic algorithms are robust and stochastic optimization techniques that overcome the mathematical complications and combinatorial explosion problems of the conventional optimization techniques. Markovian models are the state-of-the art stochastic models used in several maintenance management systems to predict the future condition for a network of infrastructure facilities. An illustrative example for finding the optimal maintenance alternatives for concrete bridge decks using field data is presented to demonstrate the feasibility and capability of the proposed approach.