Robust Optimization for Managing Pavement Maintenance and Rehabilitation

Abstract

A pavement management system should help a decision maker to select the best preservation program, decide which preservation treatment to use, and where and when to apply it to maximize the use of the available resources. One of the essential roles of pavement management is to provide a rational, cost-effective optimal funding planning and allocation strategy for highway agencies. Researchers have previously developed deterministic optimization methods for programming pavement maintenance and rehabilitation strategies. However, pavement infrastructure deterioration is a dynamic, complicated, and stochastic process affected by a variety of factors such as traffic loading, environmental conditions, and structural capacities, as well as certain unobserved factors. Ignoring these fundamental characteristics may limit the usefulness of an optimal solution. To take the uncertainties into consideration, some researchers have introduced stochastic programming techniques into pavement maintenance management. However, difficulties in characterizing the distribution of data and the substantial computational challenge have compromised the practical application of those techniques. A project-level robust optimization method for maintenance budget planning to overcome these difficulties is presented. The solutions from this proposed method are computationally tractable and not overly sensitive to any specific realization of the uncertainties. An application of this method is demonstrated by using long-term pavement performance data collected during the past 20 years, yielding promising preliminary results.