Abstract
Traditionally, bridge management systems were designed using a Markov chain decision model. Based on the analysis of 15 years of bridge inspection data, we apply the gamma process instead. After extracting all relevant information, enough data was collected on the condition paths of elements to build a deterioration model. The element conditions follow a time period in full condition then start deteriorating. We consider a random variable for the last time the condition was observed to be 100%. We consider the stochastic deterioration process that follows. The amalgamation of the two part process through probabilistic arguments creates a new stochastic process. The novel stochastic process characteristics are derived through the data to provide a predictive model for the element, bridge and network conditions. We showcase a software solution for bridge network condition assessment, monitoring and prediction.