Three-machine scheduling: source model and practical boundary

Family 124. Edition: 9 October 2026. Source metadata. Revision fd4aeeb2ee4fc729c18d98444fed42fd0529eeeb.

Problem and potential new use

A bounded model/solver assurance adapter can distinguish a feasible schedule, exact tiny optimum, deadline answer and source-machine guarantee. The seven principal TeX argument files were read, including the huge finite-description enumeration and the full invariant proof text. This is source-text review, not independent mathematical acceptance.

Applicability and commercial boundary

Nonempty explicitly listed DAG, unit nonpreemptive jobs, exactly three identical machines, no release/communication/eligibility/extra-resource constraints. The source exponent 150020 makes no practical performance claim. The selected 100-line Comparator interface, 14-line configuration, 15-line scope note and 90-line solution entry point were read; imported MachineBridge and proof closure remain unreviewed/uncompiled.

Business decision

Prototype a bounded assurance module within the shared evidence platform; defer a direct production scheduler. All eight generated controls already have matching OR-Tools optima. No buyer value, delivery cost or profitability is established. Updated dossier.

Concrete validation

The independent slot-assignment oracle agrees on all 1,099 topologically ordered DAG edge sets through five jobs. Nine-job heuristic optimum gap and six-job synchronization controls make evidence distinctions concrete. 3,346 finite controls and 40 existing-solver controls passed. Tiny DP is exponential and capped, not the source backend.

Next verification action

Measure recurring workflow value; inspect practical extracted machine/descriptor alternatives separately. Preserve source quantifiers and unknown budget states. Full proof/import/program and whole manuscript review acceptance remain pending.