Linear optimization models for long-term investment planning in a pavement network

Abstract

During the period 2004 - 2006, evaluations executed by the National Laboratory of Testing Materials and Structural Models (LanammeUCR) at the University of Costa Rica indicated that there was no a significant improvement with the road network. This conclusion leaded to establish this research based on the principles of Asset Management. A long term strategic investment plan (20 years) was developed for pavement structures for the maintenance district 1-9 (Heredia) in Costa Rica.
Several investment scenarios were estimated, including the budget assigned by the government for 2007.
For the pavement structures it was used inventories such as visual distresses, roughness (IRI), deflection data (FWD), traffic (AADT), layer and material characterization data. A Pavement Condition Index (PCI) was estimated based on deflections and roughness. Also it was used deterioration curves, selection of treatment windows, and economical analyses.
By using Woodstock Linear Optimization tool (Remsoft Inc.), several investment plans were developed at the strategic level. Some of these scenarios included different annual budgets, and also the budget required to eliminate the bad roads in 15 years was estimated. The solution includes the type of treatments that need to be applied in specific locations on a year basis and its costs. These results follow the Asset Management principle: to maximize the investment and allocate the resources at the right time, at the right asset, at the right moment, with the objective of maintaining a good level of service.
As a result, by using long term investment plans more effective decisions can be taken to recover our assets efficiently. It also can be used for accountability and request of budget. Finally, the key success to implement an Asset Management System is to address the structure of the organization, human resources and technology.