improved approach to incorporate uncertainty using Bayesian networks
The act of creating project schedules is often hampered by the difficulty involved in planning against uncertainty. This paper examines an approach to generate project schedules that incorporates risk, uncertainty, and causality via Bayesian Networks (BN). In doing so, it explains the issues involved in mitigating a project schedule's uncertainty and summarizes the methods now commonly used to develop project schedules; it lists several causes of scheduling uncertainty and describes the way that three techniques--PERT, CCS, and MCS--address uncertainty. It overviews the literature on managing uncertainty and using the Bayesian approach to handle uncertainty. It then outlines--using the CPM methodology and notations--a new approach that uses BN to schedule project activities. It notes the approach's benefits and its effect on activity durations, showing how it maps a BN from a CPM network.