Improvement of decisions about project risk management based on bayesian network

Abstract

The project risk management is a hot topic in the field of current management. As to the problems of lacking data to construct the decision model and lacking quantitative ways to inspect model,this paper sets about from the Bayesian network and puts forward a method of constructing Bayesian belief network with chain inference method, diagnosis inference method and the guidance of experts under the condition of lacking data. The network are not only used to display project risk directly in the form of graph but also used to calculate local risk and overall risk. In addition, the paper has discussed the way of using Bayesian belief construction algorithm to check out the similarity of the network structure and sample in a quantitative way combined with B-Course tool. In order to save a large amount of time of testing, the paper also optimizes the traditional mathematical Bayesian algorithm with the method of compression. Thus the decision maker can improve the utilized efficiency of time. Finally, this paper has summarized the advantages and disadvantages of this new method and provides a different way of thinking on decision-making of risk management.