Adaptive optimisation methods in system-level bridge management

In infrastructure management, the slow accumulation of facility condition data and incomplete understanding of the deterioration process can lead to inaccurate deterioration models and subsequently, suboptimal maintenance, rehabilitation and reconstruction decisions and increased system costs. State-of-the-art bridge management systems (such as Pontis) address this problem by successively updating deterioration models over time, by using a class of adaptive control methods known as the certainty-equivalent control (CEC). In this paper, we demonstrate that CEC does not necessarily guarantee improvement in deterioration model accuracy or savings in system costs for realistic heterogeneous bridge systems undergoing non-Markovian deterioration. We further apply an alternative, the open-loop feedback control (OLFC), and demonstrate its superiority to CEC in a simulation study, in which we consider a planning agency managing a system of facilities with limited prior knowledge of the deterioration models over a designated planning horizon. We show that OLFC improves model accuracy and reduces system costs. We focus our discussion on bridge decks, the component of bridge structures that undergoes the fastest deterioration, but the methodology presented in this paper is applicable to all bridge components.