A Hybrid Pareto Frontier Generation Method for Trade-Off Analysis in Transportation Asset Management

Trade-off analysis, one of the key elements of transportation asset management (TAM), helps decision makers to not only quantify how different resource allocations affect system performance but also investigate the trade-off relationships between cost and performance measures and between different performance measures. In the fast-growing field of TAM, researchers are beginning to quantify the trade-offs among the performance measures. A successful quantification of these trade-offs is important for the practice because the highway agency decision makers responsible for project selection have long been stymied by a lack of knowledge of the extent to which different decisions lead to different performance trade-offs. In a follow-up to recent research efforts that have sought to address this issue, this article presents a further improved methodology to facilitate the analysis of trade-offs in TAM. First, a general multiobjective optimization framework for TAM is established. ∗To whom correspondence should be addressed. E-mail: labi@ purdue.edu. Next, the developed hybrid method is presented, which is then shown to quicken the generation of Pareto frontiers for the purpose of conducting trade-off analysis. Using a case study, the proposed hybrid method, which was implemented on a computer to generate Pareto frontiers, ultimately converged faster and generated betterdistributed Pareto frontiers compared to the NSGA II method of Pareto frontier generation in the TAM context. Using the generated Pareto frontiers, trade-off analysis between cost and performance measures and between performance measures are demonstrated in this article.