Risk analysis using Monte Carlo simulation and Bayesian networks

The management of a global activity that has individual tasks as its components is very difficult, because an unexpected interruption in any individual task can create extra costs or even disrupt the whole activity. To resolve this problem, this work presents the development of a decision-support tool using Bayesian networks (BN). Our research illustrates how to model the relationship between the total time of a process and the time of the individual tasks selected as relevant. We use a Monte Carlo simulation to construct dynamic scenarios on the BN which allow us to track and manage a global activity. The BN is useful because the activities have random characteristics and the information about individual tasks can be propagated throughout the global activity scenario and associated with costs. This offers the administrator a tool for proactive task management and risk reduction.