Improving pavement networks through performance-based planning with optimal treatment strategies and management policies

Abstract

Performance-based planning (PBP) is an efficient way to improve pavement networks. It is the practice of using data from pavement management systems (PMSs) to support analyses on the predicted network performance based on available budgets, treatment strategies and management policies. PBP involves the collection and analyses of PMS data, pavement deterioration prediction, budget allocation, the selection of treatment strategies, and the promotion of appropriate pavement management policies.

This dissertation provides a comprehensive framework for PBP. First, it focuses on the development of a pavement deterioration prediction model and a budget allocation model. A weighted-output neural network model is proposed, which can predict multiple pavement condition metrics simultaneously and incorporate their correlations into the prediction process. During model training, each condition metric is assigned a weight to reflect its relative importance. When the weights equal to those in the formula for a multi-condition metric pavement condition index (PCI), the prediction performance for PCI is optimal (13% lower mean squared error than optimal, single-output models). In terms of the budget allocation model, a probabilistic treatment path dependence (PTPD) model has been proposed. This model incorporates uncertainties of both treatment cost and pavement deterioration, and evaluates a treatment by considering benefits of both the evaluated treatment and its following actions. Compared to a conventional benefit cost ratio model, PTPD can deliver equivalent pavement network performance with an annual budget that is 10% less.

Most existing research on PBP focuses on improving allocation decisions through changes in the allocation algorithm without considering the consequences of how optimization analyses are framed. In this thesis, both the environmental and economic performance of a pavement network are evaluated for different framings of the problem. Specifically, framings in the form of different treatment strategies that consist of treatment materials, treatment types, and evaluation period are considered. Results show that the proposed strategy that uses multiple materials (both concrete and asphalt), an increased number of treatment types, and a long evaluation period could both reduce greenhouse gas emissions and improve pavement network performance. Finally, this thesis explores the potential impact of different federal or state policies regarding PBP. Three pavement management policies are proposed, including flexible decision-making, long-term planning, and market diversification. Model results suggest that incorporating these policies for the whole U.S. pavement network (compared to a business-as-usual scenario), could reduce total excess vehicle fuel expenditures from 2017 to 2050 due to poor road conditions by 28% or about 62 billion dollars. All states can benefit from the proposed management policies.

These research findings can help transportation agencies improve their performance-based planning for pavement networks within a limited budget. In addition, this thesis also provides insights for federal or state agencies regarding the value of key policies to improve pavement networks and to reduce greenhouse gas emissions due to poor road conditions.