A Multiply Connected Belief Network approach for schedule risk analysis of metropolitan construction projects

Metropolitan construction projects may be distinguished from other types of projects due to their specific nature. Issues with regard to purchase of property obstacles, a lack of accurate layout drawings for some underground facilities and, therefore, following probable changes in designs make it vital to have specific studies on risk analysis of such engineering projects. However, conventional methods offered in literature of risk analysis fail to consider the interactional mechanisms of risks and their simultaneous occurrence. This paper exploits the Multiply Connected Belief Network approach in MATLAB environment to analyse the simultaneous occurrence of risk factors on schedule delay of a highway project. The method of inference is through Junction Tree Algorithm due to existence of multiply connections in the developed network of the project. The findings for schedule delay were tested in a different space using Monte Carlo simulation and had satisfactory results.