Addressing Uncertainties of Performance Modelling with Stochastic Information Packages – Incorporating Uncertainty in Performance and Budget Forecasts

A large volume of data is collected world-wide to feed pavement management systems (PMS). The data is typically condensed to characterize pavement sections or smaller sub-networks by using statistical measures mostly averages. In this process valuable information is lost, thus increasing the likelihood of providing inaccurate or in some cases misleading answers. The pitfalls of using averages can be avoided by utilizing the full data set and treating each data set as an entity or stochastic information packet (SIP). Modeling with SIPs means that the input as well the output of the modeling is a distribution as opposed to the singular outcome of deterministic models. The resulting distribution allows determination of the probability of the outcome besides its predicted value. Budget and condition forecasts therefore may include not only the future condition and budget requirements, but their reliability and consequently the level of associated risks. Managing agencies and contractors may choose the budget scenario best reflecting their level of risk acceptance or tolerance. Modeling with SIPs builds on deterministic models by expanding their outcomes into full distributions. Working with arrays (SIPs) requires using a novel approach that is described and illustrated in the paper.