Simultaneous Modeling for Pavement Performance Prediction and Optimal Resource Allocation: Goal Programming Approach with an Embedded Markov Decision Process

摘要

Many studies have used Markov chain method for prediction of pavement deteriorationcondition, while separating the prediction phase from resource allocation planning process. Inthis paper, a new model is developed for simultaneous pavement condition prediction andresource allocation optimization based on the Markov chain method. This model accounts forvarious rehabilitation scenarios that may be used to improve pavement condition while at thesame time allocating available resources.Different states of pavement deterioration at network-level correspond to different pavementcondition statuses (PCS) that impress different importance for the decision-makers. Each part ofthe network in each status has different values compared to other statuses. In this research,separate objective functions are developed to determine what proportions of the network fall intoeach status. This model is suited for strategic network-level resource allocation planning. Theproposed model is applied to an example network to illustrate the applicability of the model.