Modeling formalism and notifications in project management

Abstract

Not every project needs to be planned and not every missed due date requires a notification. Also, the costs of the project managers themselves are not always negligible. Here, we propose models to clarify the appropriate level of planning and information flows among project participants. Ideally, a given task becomes available to start at exactly the time when the last of its predecessors completes. If it becomes available earlier, then expensive resources associated with that task are idled. If it becomes available later, then the project may be delayed, and contract penalties may be incurred. Specifically, our work studies (a) what planning and information, i.e. notifications, the project manager should supply to the owner of the task about the time it should be ready to start, and (b) when that information should be given. We describe several notification policies: no baseline schedule, a baseline schedule, and single or multiple notifications. Our work provides optimal policies for special cases of these problems, and computationally validated heuristic policies for intractable general cases. A statistical analysis identifies the key factors that influence policy performance and robustness, thereby providing valuable practical insights to project managers.