A Pavement Maintenance Policy using Markovian Decision Model with Discounted Roadway User Cost: A Case Study of Washington, DC

ABSTRACT

Due to the rapid deterioration of today’s pavements and its effect on the flow of traffic and environment, there is a need to investigate the effect of pavement distress on roadway, hence identify effective maintenance stratagems. Forecasting the rate of deterioration of the pavement is very uncertain and modeling uncertainty needs at least some kind of optimization technique such as the Markov decision modeling process. Pavement is one of the most important elements in transportation infrastructure system which makes it also very important to be kept in good condition by executing periodical maintenance. An effective pavement management system (PMS) must be considered to achieve a good maintenance system. Due to the need and limited availability of funds allocated to projects and its environmental impacts such as pollution and traffic congestions, the issue of deciding whether or not to do maintenance of the highway infrastructure always arises.

Alternatively, the smoothness and/or roughness of the pavement in terms of the condition, can be affected by vehicle speeds. With pavement roughness affecting vehicle speeds hence negatively affecting the ride quality of the road users, there is a need for routine maintenance of the roadway pavement in order to maintain its best quality and purpose. The objective of this study is to investigate the correlation between vehicle speed and the pavement roughness, hence the development of an optimum road maintenance policy.

The pavement roughness measurement used in this study is the International Roughness Index (IRI). IRI data were obtained from the District Department of Transportation (DDOT). Speed data were also obtained from District Department of Transportation (DDOT). The optimum road maintenance policy was developed using effective data mining techniques. The data mining techniques included transition probability matrix that was developed for different states of road surface distress, their associated net economic value, and a markovian decision model that was formulated as a method of solution for the policy. There were two steps involved in developing the optimum road maintenance policy: (1) verifying that the IRI values have an impact on vehicle speeds of a particular road segment and (2) actual development of the policy improvement model. The results from the analysis indicated an increase in the mean speed value from roads in severely poor condition to roads in either poor, fair or good condition and therefore a decrease in cost. The findings of this research and the development of a maintenance policy improvement model could help in a cost-effective way of maintaining the road and also help agencies to properly allocate roadway maintenance funds in the intention of reducing the economic and environmental impacts associated with pavement roughness.