Material diversification in pavement management : a technique to proactively deal with an uncertain future

Abstract

Pavement management systems are important tools that planning agencies depend upon for the effective maintenance of roadway systems. Although uncertainty is an inherent trait of these systems, the current approaches generally exclude its consideration from the analysis. Consequently, decision-makers disregard opportunities to embed sources of flexibility that may be advantageous to deploy if future conditions unfold differently from expectations. One potential source of flexibility available to planners is the incorporation of a broader range of paving materials and designs as part of their pavement preservation strategy. More specifically, this thesis hypothesizes that the inclusion of concrete-based maintenance alternatives by an agency can act as an insurance policy that protects planners at moments of spiraling costs for other paving commodities. To test the hypothesis set forth, this dissertation develops a stochastic simulation model that incorporates uncertainty as it relates to roadway deterioration and the future cost of maintenance actions. Its greedy heuristic algorithm addresses the inability of the current methods to (a) account for the heterogeneous (e.g., material, design, traffic) nature of pavements (b) scale for the type of real-world contexts that planners intend to use pavement management systems and/or (c) allow decisions to be made sequentially over time. The algorithm provides a high fidelity solution that generally falls within 2% of the global optimum for low-dimensional and deterministic problems. Subsequently, the model is applied to the Commonwealth of Virginia's interstate system, whose department of transportation (VDOT) traditionally only maintains their pavements with asphalt-based technologies, to minimize traffic-weighted roughness over a 50-year analysis period. A comparison of the solution for Virginia demonstrates that the DOT could achieve its desired performance goals, on average, at a cost reduction of 10% by incorporating multiple paving materials as part of their pavement management strategy. Results from the simulations indicate that much of the expected benefit from the concrete-based designs stems from their ability to mitigate poor performance at times where asphalt prices are significantly higher than expected. These results suggest that the benefit from incorporating a larger range of paving materials and designs by a planning agency could be much higher than agencies realize using the current deterministic approach for pavement management.