Abstract:
Artificial Neural ,Network (ANN) has been used for risk analysis in various applications such as civil
engineering, financial, facilities management and so on. However use of ANN has become extremely difficult
when the problem is complex when handling large number of variables. Ensemble network architecture is
proposed to overcome such difficulty, by combing individual "expert networks " thai learn small parts of the
problem. In this research, ANN was used to analyze risks in maintainability of high-rise buildings. Analysis of
maintainability risks of a building involves a large number of variables as it consists with number of
components such as roof facade, etc., Therefore use of a single neural network has become impossible due to
small set of data from less number of high-rise buildings in Sri Lanka. Therefore, ensemble network
architecture was used in this research. The results showed that ensemble network has performed well in
solving complex problems (i.e. building), by decomposing the task of the problem into its sub levels (i.e.
components).
