Probabilistic deterioration prediction and cost optimization for community buildings using Monte-Carlo simulation

Prediction of deterioration of the community buildings can be complex due to the large number of elements and influencing factors. High variability of the condition data adds to the complexity and a deterministic approach often cannot be used to derive realistic deterioration progression curves. Probabilistic methods have been explored as an alternative in a current research project supported by six local councils in Victoria. A Markov process has been calibrated with the building condition data acquired from the project's partnerc utilizing a non-linear optimization technique. A Genetic Algorithm model with a Monte Carlo sampling method has been used for the convergence of the results. The paper also presents a cost optimization method using a Monte Carlo random optimization technique to assist the decision making process of the asset managers in the field. A software program developed for the project integrates the process with a comprehensive asset management data-base capabilities and a user-friendly interface to facilitate the industry requirements.