AACE PSP study materials
AACE PSP study resources

The Logic That Wasn't There

AACE PSP · Put your knowledge to work in context.

All case studies

The Logic That Wasn't There

A scheduler is building the CPM network for a 9-month plant expansion. The team provides the following logic: (a) foundations must finish before steel erection; (b) steel erection must finish before roof decking; (c) roof decking must finish before interior build-out; (d) interior build-out must finish before punch list. The scheduler also learns that the owner's equipment vendor will install process equipment "sometime after the roof is on," and that the electrical rough-in "follows the steel." No durations are given for the vendor equipment install, and the electrical contractor says rough-in is "about 6 weeks, depends." The scheduler is tempted to leave the vendor equipment install as a milestone with no predecessors and set the electrical rough-in as a fixed 6-week duration with a start date constraint. The project director wants a schedule that "won't get us in trouble at the first update."

1. Identify the scheduling defects in the proposed approach and explain why each will cause problems at the first progress update.

Read the worked answer

Defect 1 - Vendor equipment install with no predecessors (a dangling activity). An activity with no logic ties will float freely and can never be shown as driving. At the first update it will appear either artificially early or artificially late, and the critical path will be wrong. It must be tied: equipment install should be a successor to roof decking (or roof completion) with a defined lead time, and a predecessor to whatever it enables (e.g., commissioning, interior finish in equipment areas).

Defect 2 - "Sometime after the roof is on" is not logic. Convert it to a defensible relationship: equipment install starts after roof decking completes (finish-to-start) with a stated lag for rigging access. If the vendor's lead time is unknown, model it as a procurement activity with a best-estimate duration and a documented range, not as an open milestone.

Defect 3 - Fixed 6-week duration with a start date constraint. A date constraint (start-no-earlier-than) overrides logic and hides float. When steel erection slips, the constrained electrical activity won't move, so the schedule will show a false critical path and the update will mislead. Model electrical rough-in as a start-to-start successor to steel erection with an appropriate lag, and let the duration be an estimate flagged for refinement-not a constraint.

Defect 4 - "Depends" durations left unquantified. Unquantified durations become schedule risk that surfaces as surprise slippage. Use a documented estimate with a range and identify it as a risk.

Why it matters at update: constraints and dangling logic make the schedule not respond to actual progress-the classic failure that "gets you in trouble at the first update."

2. What logic and modeling standards should the scheduler apply before issuing the schedule?

Read the worked answer

Apply these standards:

  1. Every activity has a predecessor and successor except the project start and finish milestones. No open ends, no dangling activities.
  2. Use relationship types deliberately: finish-to-start for hard physical sequence (foundations → steel → roof → build-out); start-to-start with lag for overlapping trades (steel → electrical rough-in); avoid finish-to-finish unless truly warranted.
  3. No date constraints unless contractually mandated; if a constraint is unavoidable, document it and its schedule effect. Prefer logic over constraints.
  4. Model procurement and vendor work as real activities with estimated durations and identified leads-not milestones with no logic.
  5. Quantify durations with a basis (estimate source, range, confidence); flag uncertain ones as risks.
  6. Run a logic check (open ends, lags, constraints, negative float) and a critical path review before issue; confirm the driving path matches the team's physical understanding.
  7. Document the schedule basis: assumptions, lags, and estimate confidence, so the first update can be interpreted correctly.

Judgment point: the exam rewards recognizing that a schedule "that won't get in trouble" is one whose logic responds truthfully to progress-constraints and dangling activities are the enemy of that.