A Methodology for the Selection of Pareto Solutions Developed for a Multi-objective Pavement Maintenance Program (05-1749) - MP-04
Matthew R. Horrocks, Horrocks Engineers
Piya Chootinan, Utah State University
Anthony Chen, Utah State University

Recent advancements in the development of optimization techniques for pavement maintenance programming have been focused on multiobjective optimization with evolutionary algorithms. Although multiobjective optimization is a powerful method of developing multi-year pavement maintenance plans at both the network and project level, researchers have neglected to develop effective methods for selecting the Pareto optimal solutions generated during this optimization. Due to the rule based nature of optimization, it is infeasible for optimization to generate solutions that fulfill all needs of the pavement network, requiring different approaches to alternative evaluation and decision analysis in pavement management. This paper addresses the issue of Pareto optimal solution selection by utilizing Simple Multi Attribute Rating Technique (SMART) analysis. SMART analysis is a multi-criteria decision analysis method intended for the structured selection of multiple alternatives using multiple evaluation criteria. The formulation and development of this analysis are described and demonstrated using a numerical example problem in which the Pareto optimal solutions of a case study pavement maintenance programming analysis developed using a multiobjective optimization, was evaluated.