Bridge Network Maintenance Optimization Using Stochastic Dynamic Programming

Abstract

This paper presents a stochastic dynamic programming (DP) procedure for multiobjective optimization of bridge network maintenance planning that involves a group of existing highway bridges with various remaining service lifetimes and different reliability importance factors to the bridge network. The complex multiobjective optimization problem is solved by using a two-phase DP approach. The Phase I problem consists of identifying the optimal maintenance plans for individual bridges that have minimum life-cycle maintenance costs, while satisfying both condition and safety requirements for a targeted service lifetime period. This problem is solved by using a specific DP optimization algorithm along with Monte Carlo simulation. The Phase II problem is to rationally allocate the limited annual maintenance budgets in such a way that the identified optimal maintenance plans for individual bridges can be satisfied for as many bridges as possible. A single-objective formulation derived from multiple attribute utility theory with weight assignment from reliability importance factors is developed. This is solved by a binary integer programming algorithm. The ultimate goal of this study is achieved in terms of finding the most efficient combinations of available maintenance actions applied to all bridges in a highway network.