The CPM algorithm’s ability to develop bounding early/late dates using activity durations and logic was a new scheduling paradigm that has sustained CPM as the scheduling method of choice since 1957. However, in Monte Carlo probabilistic scheduling, CPM develops only the early completion distribution curve.
This keynote will: 1) reveal how the GPM probabilistic scheduling algorithm, by allowing activities to float in every iteration, develops both the statistical early and late schedules, unveiling early/late bounding completion distribution envelopes from which practitioners can infer more realistic 80% confidence level completion aka P80 dates; and 2) question continued reliance on CPM risk assessment tools that promote optimistic P80 dates. The hoped-for outcome is a path toward more realistic schedule risk analysis results.