Portal de Periódicos da CAPES

Method for Generating Multiple MRR Solutions for Application in Cost-Leveling Models

2016; American Society of Civil Engineers; Volume: 23; Issue: 3 Linguagem: Inglês

10.1061/(asce)is.1943-555x.0000351

1943-555X

Yoojung Yoon, Makarand Hastak, Kyuman Cho,

Concrete Corrosion and Durability

Cost-leveling models have been introduced recently to minimize the fluctuations in annual maintenance, repair, and rehabilitation (MRR) costs within a multiyear infrastructure asset-management plan. However, these models were found to produce limited results because they ignored the impacts on all the preceding/succeeding MRR activities in a multiyear plan when the initially identified implementation time of one MRR activity was adjusted for cost-leveling purposes. This paper proposes a method to generate multiple MRR solutions at the facility level for existing cost-leveling models at the network level. The method utilizes two concepts as follows: (1) the fast elitist nondominated sorting genetic algorithm (NSGA-II) to generate a Pareto optimal set of feasible MRR alternatives, and (2) the pseudoweight vector approach with the Euclidean distance and a confidence interval to select multiple MRR solutions from the Pareto optimal set identified by NSGA-II. The proposed method is applied to a sample concrete bridge deck as a case study in this paper; the results show that the proposed method was able to provide multiple MRR solutions having various profiles in terms of the intervention times of the MRR activities and annual MRR costs. Consequently, the proposed method advances the theoretical framework of existing cost-leveling models by combining four approaches to provide multiple MRR solutions at the facility level: (1) NSGA-II algorithm, (2) pseudoweight vector approach, (3) Euclidean distance, and (4) a confidence interval. In addition, it is expected that the proposed method can be well applied to existing multiobjective optimization models to generate a set of nondominated good policies for infrastructure asset management.