Software Project Risk Management Using Bayesian Belief Nets (BBNs)

Abstract

Risk management has been considered as an effective technique to cope with inherent uncertainty associated with software development. Current software risk management mainly relies on subjective judgment of project managers. In such subjective approach, decision-making process is human-intensive and opaque. This paper presents a systematic approach to assist software risk management using BBN (Bayesian Belief Network). It converts the opaque decision-making process in risk management into a visible and repeatable process. Our method uses BBNs’ causal dependency and prediction function to assist managers to perform risk analysis and resource adjustment. We have implemented the proposed algorithm in a simulation environment. Our results showed that this BBN-based approach could effectively enhance probabilities of project success.