An optimization methodology is presented for program planning and budget allocation involved in network-level pavement management. It consists of three major components: (a) characterization of pavement condition into discrete states; (b) specification of treatment options for each pavement state, and (c) application of a linear programming technique for constrained optimization and development of the multiyear pavement program. Pavement sections having similar characteristics are classified into states that are defined on the basis of distress and nondistress factors. The network condition is represented as lane miles of pavement distributed among various states. Several treatment options are specified for each state; they are based on the information incorporated in the state definition. For each treatment applied, the time for a complete state transition (or increment) to occur is predicted from historical data and empirical knowledge. A linear program is formulated to model interactions between economic and engineering factors in an effective manner. It enables decisions about the type of treatment, timing, and magnitude of work to be made simultaneously. Both project- and network-level constraints can be imposed to develop a pavement program that meets specified requirements on condition and budget. The developed methodology has been implemented as part of the New York State Thruway Authority's pavement management system. An example is presented to illustrate the methodology and its usefulness to conduct variational analysis. It is concluded that the methodology can be applied to develop an effective multiyear pavement program and ensure optimal budget allocation for the entire network. (A)
Abstract