Analytical Framework for Optimizing Pavement Maintenance

Abstract

The problem of scheduling maintenance for pavements in an optimum fashion has been approached in a variety of ways by researchers and practitioners. However, the Markov decision process has found very limited use despite the fact that cumulative damage is readily modeled by a Markov chain and that a wealth of immediately applicable theoretical results exist in the literature. The solutions are known for a variety of problems involving inspection, repair, and replacement, making it possible to solve directly for an optimal policy in the form of a control law. This paper reviews some of the relevant theoretical results in order to call them to the attention of civil engineers involved with pavement management systems.