methodEnglish

Model aggregate duration or a CPM network

Choose an honest schedule resolution, validate predecessor logic, and distinguish activity impacts from project-finish impacts.

EnglishEspaƱolFranƧais

Documentation follows the current system language. Missing translations fall back to the page's canonical language.

Model aggregate duration or a CPM network

Aggregate mode models total project duration as one baseline value with optional base uncertainty. Use it when a validated network is unavailable and state that it cannot reveal critical-path switching or activity-level propagation. Network mode uses activities, base durations, predecessor IDs, and optional duration distributions. Every simulation iteration samples activity durations and recalculates the critical-path finish.

The network must be a directed acyclic graph. Duplicate or broken predecessor references, self-links, and cycles block simulation. Verify that activities cover the declared project boundary, use one duration unit, and exclude schedule reserve and risk allowances. This version does not model working-day calendars, resource leveling, crews, lags, or resource-loaded execution; document those omissions instead of implying they are present.

Place impacts correctly

An activity-targeted schedule impact changes a named activity before the critical path is recalculated. A project-finish impact is added after network calculation. Choose the target based on mechanism, not convenience. A delay affecting several activities may require a better network or explicit scope decision; duplicating one impact across activities inflates exposure.

Review the deterministic baseline finish before adding uncertainty. Unexpected path behavior is a schedule-logic issue, not something Monte Carlo should conceal. Report critical-path switching cautiously: it shows which sampled paths govern modeled finish, not a promise about field execution. If schedule logic is immature, preserve that warning even when cost simulation remains valid. Next, define dependencies and correlations. ## Interpretation discipline

Keep inputs, derived results, and human decisions visibly separate. Every material number needs linked evidence or an explicit assumption flag and basis. Unknown is not zero. A percentile is conditional on this model, scenario, seed, engine version, and source version; it is neither a guarantee nor an approval. Any material model edit makes an earlier result stale.

References

GAO Cost Estimating and Assessment Guide (GAO-20-195G)

A reference for credible cost estimates, risk and uncertainty analysis, sensitivity, documentation, and management review.

https://www.gao.gov/assets/710/706933.pdf

NASA — Program Planning and Control Guidance Documents

Official cost-estimating guidance, including material on cost-risk analysis, distributions, correlation, and joint confidence levels.

https://www.nasa.gov/ocfo/ppc-corner/ppc-guidance-documents/