Integrated Pavement Management System with a Markovian Prediction Model

TLDR

The system uses a discrete‑time Markov model to forecast pavement deterioration, incorporating maintenance actions, and employs two decision policies—one optimizing a nonlinear objective under budget limits and another minimizing cost while meeting condition targets—selected via random or worst‑first approaches and solved with penalty‑function and uniform‑search optimizers. The integrated pavement management system delivers an effective decision‑making tool for planning and scheduling maintenance and rehabilitation, supported by two major decision‑policy options.

Abstract

An integrated pavement management system has been designed to provide the pavement engineers with an effective decision-making tool for planning and scheduling of pavement maintenance and rehabilitation (M&R) work. The developed system applies a discrete-time Markovian model to predict pavement deterioration with the inclusion of pavement improvement resulting from M&R actions. An effective decision policy with two major options has been used. The first option optimizes a generalized nonlinear objective function that is defined in terms of proportions of pavement sections in the five deployed condition states, and is subjected to budget constraints. The second option minimizes M&R cost which is subjected to preset pavement condition requirements in terms of state proportions at the end of a selected study period. The system applies two approaches for the selection of pavement project candidates. The first approach is based on random selection of pavement sections within the same condition state, while the second one relies on worst-first selection within the same condition state. The optimization process is performed using two different optimization methods which are the penalty function method and uniform search method.