A general iterative approach for the system-level joint optimization of pavement maintenance, rehabilitation, and reconstruction planning

The objective is to minimize the total costs incurred to both the highway users and the pavement management agency. We propose a Lagrange multiplier approach together with derivative-free quasi-Newton algorithms to solve the problem for two scenarios: i) with a combined budget constraint for all the treatments; and ii) with one budget constraint for each treatment. The system-level solution approach has the following merits: i) it can be applied to problems with any forms of segment-level models for user and agency costs, deterioration process, and treatment effectiveness, given that the solution to the segment-level problem is available; ii) under the combined budget constraint, it ensures that the optimality gap of the system-level solution is bounded by a term that depends upon the optimality gap of the segment-level solutions; and iii) it exhibits linear complexity with the number of segments. At the segment level, a new maintenance effectiveness model fitted on empirical data is proposed and incorporated into the MR&R optimization program. A greedy heuristic algorithm is developed, which greatly reduces the computation time without compromising the solution quality. Combining the system- and segment-level models and solution algorithms, we examine a batch of numerical cases. The results show considerable cost savings from the incorporation of maintenance, and from jointly optimizing the use of a combined agency budget. A number of managerial insights stemmed from the numerical case studies are discussed, which can help highway agencies formulate more cost-efficient MR&R plans and budget allocation.