Assignments of Pavement Treatment Options: Genetic Algorithms versus Mixed-Integer Programming

Abstract

Assigning specific maintenance treatments is an important process in a pavement management system (PMS). A decision-making method to support this process should be based on an objective of optimizing the service life and cost of each treatment, which becomes a multiobjective process when applied to a road network. This paper explored the expected accuracy rates of network treatment options through a multiobjective optimization methodology which utilized genetic algorithms (GAs) and mixed-integer programming (MIP). This paper demonstrated the application of GAs and MIP based on the common indicators of distress for evaluating pavement condition (rutting, raveling, potholes, cracks, and roughness). The results indicated that the proposed method is capable of effectively assigning pavement maintenance while considering optimal service life and minimal cost.