Schedule Risk and PERT in Undergraduate Capstone Projects
Michael Van Hilst, Reginald Parker · 2024
Abstract Capstone projects often do not go according to plan. They fall short and require heroic effort. We looked at the kinds of risks generally discussed in the literature on risk management, both in industry and for capstone projects, and compared them with the kinds of risks commonly encountered in student capstone projects, especially software intensive capstone projects. Students are inexperienced, they face knowledge gaps, there is little or no budget, and the schedule is fixed – it's the end of the semester. The project management model taught in Sommerville, Wysocki, PMBoK, Prince2, SEI, and INCOSE are based largely on the CPM practices of DuPont. DuPont created CPM to build chemical plants. They knew how long every task would take. They had already built more than a hundred. Conversely, PERT was developed for the Polaris submarine. No one had ever built a submarine to launch missiles. They developed practices to identify the biggest risks and address them early on. Capstones are as much or more like the first Polaris submarine as the 151st chemical plant. In this paper we propose an alternative approach to risk management and project scheduling for student capstone projects. The key idea is to address risk on a per-task rather than project-wide basis. The proposal revives practices from the original PERT method of scheduling and estimation, which are still applied in industry by technical managers, but, for the most part, have been missing in the models of project development taught to students. We describe how to use three-point estimation to identify task-level risks, and then measure both schedule impact and exposure directly from the initial three estimates. We present an expanded set of strategies based on task-level risk and the principle of moving your exposures to the front, in which risks are managed, not just monitored. Finally, we discuss how to apply this approach in the context of Waterfall, Spiral, and Agile development. The PERT practices, documented in the 1958 Phase 1 report, and experienced first-hand by one of the authors, are easy to apply. They were developed at a time when they had no computers. The second author is a certified PMP instructor and a turn-around specialist for failing projects in industry. Here at Embry-Riddle, we have begun an experiment to help students manage several capstone projects, and two undergraduate research projects, involving both technical risk and significant knowledge gaps. The plan is to augment and refine, rather than replace, the existing models, while incorporating the knowledge factory view and parallel development from Lean, prototype planning from Spiral, and spikes and burn-down from Agile. We also borrow from NASA's FMEA to isolate risks in the Work Breakdown Structure. While these practices exist in industry, what is new is putting them together with per-task risk estimation in a way that can be applied by students to better anticipate challenges, and produce more realistic schedules. In this paper, we present our findings from an extensive literature review, and discuss the elements of a cohesive plan for students.