Aster Museum · decision brief for Omar Bell · 25 January 2027

Protect the 15 March opening with £24k.

Recommend S + T: hire the second fit-out crew and buy out the technician's booking. This gives a feasible operating run finishing at boundary 24, three working-day boundaries before acceptance at 27. The sponsor can approve exactly £24k. This is a proposed commitment, not an approval or a test certificate. M4 M5 M6

The least-cost on-time package is S alone, £18k, but its operating run ends exactly when acceptance starts. With original resources, the earliest feasible opening is 24 March, boundary 37: seven working days late and £14k holding cost. These are deterministic schedule results under the supplied durations and calendars. M1–M6 Calculation method

Correct the draft package report.
E7–11, F6–9 and G8–12 all claim S0 at once. Those dates cannot coexist; the draft's no-cost 15 March opening is unsupported. M7

The decision in three feasible plans

Each plan reserves the earliest permissible acceptance slot after P. “Margin” is P finish → Q start; it is not a probability of success. Open any row to inspect all eighteen activities.

PlanAccelerationP finishesQ startsMarginPublic openingHolding / totalInspect

All costs are additional £000. Total = acceleration + £2k × max(0, opening boundary − 30). No additional financial consequence of delay is quantified. M6

Interactive planning copy · ephemeral

Inspect the funded proposal or test another package

Select purchases to load a checked feasible schedule for that exact combination. Changing purchases replaces any manual trial. Nothing here purchases an offer, reserves a slot, creates approval or changes source records; reload/reset restores the recommended S + T plan.

Firm offers expire before day 0. Off-site preparation is included. M5

Scarce-resource lanes

Bars occupy [start, finish): every day of a bar must be available. M and P appear on both C0 and T because both are required together. No crew handovers. Axis labels are workday boundaries. M1 M3

Assigned workJoint C0 + T workUnavailableDashed line: selected Q start

Activity and dependency plan

Boundary 0 = Monday 1 February 2027; Monday–Friday, no holidays. Finish is the boundary after the last occupied workday, so a Friday's completion is labelled at the following Monday boundary. Successors may begin at the same boundary. “Wait” = start minus latest original predecessor finish; it includes resource/calendar/booking effects, not just idle capacity. M1 M2

IDActivityDaysPredecessorsAssigned resourceStart boundaryFinish boundaryWaitReason for wait in this plan

Test a specific move or crew assignment

This holds other activity starts fixed and checks the resulting plan; it does not re-level successors. Q and R are automatically moved to the first remaining booked slot after P. If any precedence, calendar or resource conflict remains, the opening claim is withdrawn. Restore the package plan to return to its calculated sequence.

No manual trial applied.

All 32 listed purchase combinations and calculated outcomes

Each “Open” button retrieves its complete feasible schedule. These deterministic dates assume all tests pass; none is a recorded approval.

Offers£kEarliest P finishMargin to QOpeningTotal £kAuthoritySchedule

Omar's recommendation and decisions required

  1. Seek sponsor approval for S + T, £24k, before day 0. It is within delegated authority, and pays £6k above the cheapest on-time package for three boundaries of acceptance-readiness margin. S alone remains a viable £18k fallback, with zero margin. Neither approval is recorded. M5 M6
  2. Reserve Q at boundary 27, days 27–28 (10–11 March). Then rehearse on day 29 (12 March), opening at boundary 30 (15 March). Availability is supplied; reservation and successful evidence are still required. M4
  3. Commit the actual crew sequence. For S + T: F on S0 at 6–9; E then G on S1 at 8–12 and 12–16. C0 performs I9–11, H12–15, J16–20, M20–22 and P22–24. T performs K12–15, L16–18, M20–22 and P22–24. Obtain the relevant managers' commitments to these assignments. C1 is not purchased or needed. M2 M3 M5
  4. Keep the public commitment and test evidence separate. Board approval alone can change the advertised date. A missed Q27 means the next booked slot is Q34, followed by an opening at 37 if all preceding work is ready. No overtime or partial opening is authorised. M1 M4 M6

Why the extra £6k can matter

With S alone, L cannot span T's unavailable days 17–18, so it runs 19–21. M runs 21–23, but P cannot run 23–25 because C0 is unavailable on day 24; P must run 25–27. With the T buyout, L runs 16–18, M20–22 and P22–24: the operating run finishes before C0's absence, leaving three boundaries to Q27. M3 M5

That margin is useful but cannot absorb every three-day upstream overrun: absences and non-preemption can cause jumps. It is not a guarantee.

What the cost arithmetic does—and doesn't—say

The original-resource plan costs £14k in holding charges, less than £18k for S or £24k for S + T. Buying acceleration therefore serves the strong advertised-opening commitment; the quantified holding cost alone does not justify it. M6

For one avoided missed slot only, £6k ÷ £14k = 42.9%: the extra buyout would break even if it reduced the probability of that one-slot delay by more than 42.9 percentage points. This is a conditional illustration, not a probability estimate. No duration-risk distribution or value for broken public commitments is supplied. M8

Other offers show why shorter bars aren't enough

X alone (£7k), C alone (£12k), and T alone (£6k) still open at boundary 37 in their best schedules. X + C (£19k) reaches Q27 and opens at 30, with zero margin: preassembly lets G fit at 9–12 before S0's absence, while C1 can handle J at 14–18. The combination matters, but S alone is £1k cheaper. V shortens K, yet adding it to S (£22k) leaves the same zero-margin opening plan because L, M and P's calendars still govern. Open those combinations above to inspect each sequence. M2–M5

Calculation, checks and limits

Optimal within this finite fixture: all 32 purchase combinations were enumerated. For each, the calculation enumerates every permitted whole-activity S0/S1 and C0/C1 assignment and every local resource order for E/F/G, H/I/J and K/L. M and P follow their stated predecessors and use C0 + T together. This covers 14,400 assignment/order cases across the packages.

For each fixed order, activities are placed at their earliest predecessor- and calendar-feasible boundary. Any feasible schedule induces one of the enumerated resource orders; delaying a job in a fixed order cannot improve a successor under these minimum finish-to-start constraints and fixed calendars. The selected schedule minimises P finish, then total start boundaries to choose a readable tie. Q is the earliest of 27/34/41 that P can meet; R follows Q. Non-scarce independent work is placed at its earliest boundary.

D(G) = X ? 3 : 4; D(K) = V ? 2 : 3 Each activity occupies [s, s + d), on one permitted assignment throughout. Every predecessor finishes ≤ activity start; shared-resource intervals do not overlap. Q = min {27, 34, 41} such that Q ≥ finish(P) Opening = Q + 2 + 1; readiness margin = Q − finish(P) Holding (£k) = 2 × max(0, opening − 30) Total (£k) = selected offer costs + holding

Every retained schedule was separately checked for all 18 durations and dependencies, each occupied resource day, resource overlaps, A0, and the fixed acceptance slots. The live page repeats these checks after package selection or a manual trial.

This is a feasible deterministic remaining-work plan, not a certificate that tests or acceptance will pass. The optimality claim is limited to the supplied offers, fixed durations, calendars, non-preemption, no overtime, and listed slots. Future negotiated offers, stochastic overruns, unquantified consequences and booking loss are outside the model. “More resilient” means more scheduled P-to-Q margin, not a statistically ranked risk optimum. M4 M8

Provenance files accompany this page: enumeration and independent checker · all calculated schedules · unverified self-assessment · proposed browser journeys. No external sources or network assets are used.

Original evidence and dated position

Source records remain fixed. The controls above change a hypothetical planning copy only.

M1 · Programme office, 20 January 2027

Boundary 0 is Monday 1 February; workdays Monday–Friday, no holidays. Duration d starting at s occupies s through s+d−1 and finishes at s+d. All work is non-preemptive, cannot span an unavailable required-resource workday, transfer crews midway or pause around absence. A starts at 0. Opening is permitted only after all eighteen activities, at R's finish. Advertised opening is boundary 30, Monday 15 March. No partial opening or overtime approved.

M2 · Signed remaining work, package leads, 21 January

IDActivityOriginal durationPredecessorsOriginal resource

Different non-scarce teams are independent and always available. S0, C0 and T each represent one indivisible person/crew.

M3 · Confirmed calendars, resource managers, 22 January

S0 unavailable on days 12,13; C0 on days 15,24; T on days 17,18. Otherwise available days 0–47. Required assignments cover every occupied day. Specific M5 purchases are the only substitutions. C0 must perform M/P even if C1 is hired.

M4 · Acceptance booking, external acceptance manager, 22 January

Q must occupy exactly days 27–28, 34–35 or 41–42, starting at 27,34 or41 after P. These are the only slots, all available to reserve now at equal price. Both days are mandatory. The booking does not reserve C0/T for earlier work. Acceptance and rehearsal depend on successful evidence.

M5 · Firm acceleration offers, commercial manager, 23 January

X £7k: G4→3, same resource. S £18k: S1 available days 8–16 inclusive for whole E/F/G assignments. C £12k: C1 available days 14–21 inclusive for whole H/I/J assignments, never M/P. T £6k: technician also available days17,18. V £4k: K3→2, unchanged resources/dependencies. All offers combinable, expire before day0, include completed off-site preparation. No pooled calendars or mid-task handover.

M6 · Sponsor funding/delegation, 23 January

Sponsor may approve acceleration up to £24k on Omar's recommendation; more needs the board. Holding costs are £2k per working-day boundary after30, in addition to purchases; they do not accelerate work. No other financial consequence quantified. Only the board can change the strong advertised-opening commitment; it is not a safety waiver. Using all allowance needs justification.

M7 · Initial unintegrated package report, 24 January

North E7–11, South F6–9, East G8–12 are all labelled “crew confirmed”; summary claims “opening 15 March, no extra cost”. There is no resource-levelled plan or signed acceptance-readiness date. M2–M4 are authoritative; M7 supplies no additional resource promise.

M8 · Director's decision requirement, 25 January

Every feasible option needs an auditable sequence. Distinguish opening from P-to-Q margin. Waiting for a booking can mean acceleration improves resilience without changing opening. No probability distribution of duration overrun is supplied. A future negotiation is not an available offer.

Read the complete supplied case · BPB2-03
# BPB2-03 — Museum refurbishment — produce a credible opening plan

Suite BPB2 v0.3-pilot; Foray 247, third essai.

### Scenario

Aster Museum has sold tickets for reopening on 15 March 2027. Three gallery packages share fit-out specialists, conservators and a systems technician whose outside commitments differ. Integrated acceptance has booked windows rather than unlimited availability. Delivery manager Omar Bell must replace optimistic package dates with a feasible remaining-work plan and choose an acceleration package. Small changes may merely move waiting time, while a combination that protects the acceptance slot can avoid a week of public-opening delay.

### Exact task prompt

Produce one usable in-browser answer page for the named practitioner, using the supplied fictional evidence. Keep observed facts, forecasts, assumptions and proposed decisions distinguishable, and cite source IDs beside material claims. No external research or real-world certification is required.

Prepare Omar's resource-feasible opening plan and acceleration decision. Establish the earliest feasible opening without purchases, identify the least-cost listed package that meets the advertised opening, and compare a useful more-resilient package within delegated authority. Show every activity's start/finish, assigned scarce resource, acceptance-slot readiness and any waiting time that controls the result. Explain interaction between purchases, resource calendars and the booked acceptance window; do not infer opening improvement from shorter activity bars. Recommend a funded package and the decisions required, with a feasible schedule rather than a generic critical-path argument. State any limitation of your claim of optimality.

### Supplied records

**M1 — Calendar and completion convention, programme office, 20 January 2027.** Boundary 0 is Monday 1 February. Workdays are Monday–Friday, no holidays. An activity of duration d starting at boundary s occupies days s through s+d−1 and finishes at boundary s+d; successors may start at that boundary. All activities are non-preemptive: they cannot span an unavailable workday of a required resource, transfer crews midway or pause around an absence. Advertised public opening is boundary 30, Monday 15 March. All eighteen activities must finish first; opening is permitted at R's finish boundary. A starts at 0. Other work may wait if needed. No partial opening or overtime is approved.

**M2 — Remaining-work schedule, signed by package leads, 21 January.** Different non-scarce teams are independent and available whenever needed. S0, C0 and T are the three scarce resources; each is one indivisible person/crew.

| ID | Activity | Workdays | Predecessors | Required resource |
|---|---|---:|---|---|
| A | Common service isolation/renewal | 3 | None | Services team |
| B | North preparation | 4 | A | North preparation team |
| C | South preparation | 3 | A | South preparation team |
| D | East preparation | 5 | A | East preparation team |
| E | North specialist fit-out | 4 | B | S0 |
| F | South specialist fit-out | 3 | C | S0 |
| G | East specialist fit-out | 4 | D | S0 |
| H | North objects/cases | 3 | E | C0 |
| I | South objects/cases | 2 | F | C0 |
| J | East objects/cases | 4 | G | C0 |
| K | AV/control configuration | 3 | E,F | T |
| L | Climate commissioning | 2 | G | T |
| M | Integrated case/alarm tests | 2 | H,I,J,K,L | C0 and T together |
| N | Interpretation/signage | 3 | E,F | Interpretation team |
| O | Accessibility/visitor-route walkthrough | 1 | N,H,I,J | Operations team |
| P | Full integrated operating run | 2 | M,O | C0 and T together |
| Q | External acceptance | 2 | P | Booked acceptance team |
| R | Staff rehearsal | 1 | Q | Operations team |

**M3 — Confirmed resource calendars, 22 January; resource managers.** S0 is unavailable on workdays 12 and 13. C0 is unavailable on workdays 15 and 24. T is unavailable on workdays 17 and 18. Otherwise each is available from day 0 through day 47. Required assignments apply throughout every occupied day. Subcontractors cannot substitute without the specific purchases in M5. The object conservator required for M/P must be C0 even if an additional conservator is hired.

**M4 — Acceptance booking, 22 January; external acceptance manager.** Q must occupy exactly days 27–28, 34–35 or 41–42, starting at boundary 27, 34 or 41 respectively. P must be finished by that start. These are the only slots in the fixture and all are available to reserve now, with no price difference. The full two days are required. Do not infer a free inspector at an intermediate date. The acceptance team's availability does not reserve C0/T for preceding work. Acceptance and rehearsal still depend on successful evidence; the schedule is a feasible plan, not a certificate that tests will pass.

**M5 — Firm acceleration offers, 23 January; commercial manager.** Prices are additional £000. All offers can be combined and expire before day 0; their off-site preparation is included in the quoted price and completed before its stated effect.

| Code | Purchase and exact effect | Price |
|---|---|---:|
| X | Preassemble east cases: G duration 4→3, same S resource | 7 |
| S | Second interchangeable fit-out crew S1, available only days 8–16 inclusive; E/F/G may be assigned wholly to S0 or S1 | 18 |
| C | Additional conservator C1, available only days 14–21 inclusive; H/I/J may be assigned wholly to C0 or C1; C1 cannot perform M/P | 12 |
| T | Buy out technician's other booking: T becomes available on days 17 and 18 as well | 6 |
| V | Preconfigure AV: K duration 3→2, all predecessors/resources unchanged | 4 |

A crew's activity must fit wholly within that crew's availability. Do not treat the two calendars as a pooled capacity that allows a mid-task crew handover.

**M6 — Funding/delegation, 23 January; sponsor.** Up to £24k acceleration can be approved by the sponsor on Omar's recommendation. More requires the board. Ordinary holding cost is £2k for each working-day boundary by which public opening is later than 30; it is additional to acceleration spending and does not buy faster work. No other financial delay consequence is quantified. The board alone may change the advertised opening, which is a strong commitment, not a safety waiver. Do not assume that spending the whole allowance is automatically justified.

**M7 — Package report, 24 January; unintegrated draft.** North shows E7–11, South F6–9, East G8–12, all labelled “crew confirmed”; the project summary therefore says “opening 15 March, no extra cost”. No resource-levelled schedule or signed acceptance readiness date supports that summary. M2–M4 are the current authoritative work/calendar records; M7 is evidence of the position needing correction, not an additional resource promise.

**M8 — Director's decision requirement, 25 January.** Present at least one auditable sequence for each option you claim feasible. Distinguish public-opening date from the margin between P finishing and Q starting: waiting for a booked slot can mean a useful acceleration improves resilience without changing opening. No probability distribution for duration overrun is supplied. Omar may negotiate a different future offer but may not count it in this decision as an available resource.

### Browser deliverable and completion conditions

A readable activity/dependency plan plus separate resource lanes and an opening/cost decision. Selecting purchase combinations must generate or select verified feasible schedules, flag impossible resource assignments and show P-to-Q readiness margin. At minimum retain the original-resource plan, recommended plan and resilience comparison. Moving an activity or changing a purchase must not silently preserve a now-infeasible date. Display actual scarce-resource assignments, the accepted calendar convention and authority limits. Include a clear return to the gallery.

The page must work in the supplied browser/runtime; a download button or file preview alone is insufficient. Include a clear “Return to gallery” link using the shared shell route, or ../index.html if no route is supplied. Required controls must change the dependent results, not just their appearance. Original evidence and the initial dated position must remain recoverable; hypothetical changes do not create approvals or alter source records. Optional exports are secondary.