Assessing project portfolio risk based on Bayesian network
Dujuan Guan, Keith William Hipel, Liping Fang, Peng Guo · 2014
The risk of a project portfolio is assessed using a new methodology that identifies the risk transfer in projects by using a Bayesian network structure learning algorithm to construct an interdependent network of risks. In particular, to overcome the drawback of a greedy search algorithm having a random starting structure, the mutual information between project risks is measured before executing the algorithm by computing the impact of technical interactions on project risks. It is demonstrated that the preprocessing can exclude the error and indistinctive connections, so as to reduce the search space effectively. Finally, the project portfolio risk of a practical case is assessed using this method, and the results show that the interdependent network of risks is an effective tool to reveal the risk transfer in projects and infer the value of project portfolio risk.