Abstract
A significant proportion of building investment expenditure goes to replacement expenditure for organizations owning a large building stock or portfolio. Over the years, researchers have attempted to develop asset replacement models to aid decision-making in building portfolio management, based upon' a statistical or an heuristic approach. This study attempts to use genetic algorithms to develop models for forecasting long term asset replacement strategies, aiming at smoothing fluctuations of expenditure and resource requirements, and most importantly minimizing the total maintenance and replacement costs. Scenarios are presented to demonstrate how these can be achieved. Further refinement for practical application of the models is also presented.