Plan · Location M (greenhouse · existing)

Weekly planting, 2025–2029

validated with grower

42 bays, one white cultivar, revenue at the harvest week's expected price. The model adds a contract harvest floor, prices extra weekly capacity, and tests the Women's-Day window under price uncertainty.

Baseline plan

—

4-yr total · validated

Contract-feasible optimum

—

4-yr total · ≤6 planted · ≥2 harvested / wk

Cost of the contracts

—

4-yr total vs plan without harvest floor

Weeks below minimum supply

—

baseline plan · under the committed 2 kappen/week

7th kap / week worth

—

4-yr total of extra revenue

Women's Day exposure

—

share of revenue in wk 05–10

The four-year outlook is the point of planning ahead rather than repeating last year: planting is committed months before prices are known, so the plan's value lies in what can be weighed in advance. The same horizon feeds the price-uncertainty analysis below — risks are evaluated before they arrive, not read off last season's results — and every new year of data makes those choices finer-tuned.

Year

The plan — kappen planted per week

Bars: contract-feasible optimized plan. Step line: the baseline rhythm. Orange: the estimated price calendar (€/stem) — the Women's Day peak (wk 08–09) and the autumn window are what the plan chases. Yearly view shows that year's planting weeks; the revenue readout above counts harvests landing in that calendar year.

The price of fixed contracts

Fixed delivery contracts commit the greenhouse to supplying buyers every single week — at least 2 kappen harvested, possibly 3, peak week or not. The rows compare two scenarios of the same model: planting freely, and planting under that weekly floor. The difference is the price of the commitment. (The baseline plan was drawn up without the floor and falls below it in — weeks.)

Scenario	Revenue 2025–2029	Cost of the commitment

Flexibility to respond to the market

The 6-kappen-per-week pace is a labor limit — what the harvest crew can handle. Extra weekly capacity does not mainly mean more flowers; it means freedom to concentrate work into the weeks whose harvests land on the price peaks. Each step is worth less than the last: the first extra kap serves the most valuable peak week, the next a less valuable one — and the greenhouse itself takes over as the limit, since 42 bays and fixed cycle lengths cap what can be in the ground at once.

Planning answer

—

Planning under price uncertainty · Women's Day (harvest wk 05–10)

The Women's-Day price peak is certain — how high it goes is not

tested under three price futures

Planting is committed about ten weeks before Women's Day, but the price is only known at harvest — the plan cannot wait for it. So every candidate plan is tested against three futures: prices come in strong (50% above the estimate), as estimated, or weak (40% below). The tabs below compare how each plan holds up.

Plan

How this plan holds up in each future

Revenue 2025–2029 under each price future, and what it averages out to.

Strong prices

—

30% likely

As estimated

—

45% likely

Weak prices

—

25% likely

Average outcome

—

weighted by likelihood

If prices disappoint

—

average of the worst outcomes

Revenue in WD weeks

—

share earned in wk 05–10

What this plan changes on the ground

Kappen planted per week through one Women's-Day run-up (autumn to spring). Step line: the baseline rhythm, for comparison.

All plans side by side

The selected plan's row is highlighted. Click a tab above to inspect a plan.

Plan	Strong prices (30% likely)	As estimated (45%)	Weak prices (25%)	Average outcome	If prices disappoint	Revenue in WD weeks

The percentages are the assumed chances of each future — roughly, a strong Women's Day in 3 years out of 10, a typical one in about half, a weak one in 1 year out of 4. Average outcome weighs the three revenues by those chances. If prices disappoint is the average of the worst quarter of outcomes — the number a cautious planner watches. Revenue in WD weeks is the share of total revenue earned in harvest weeks 05–10: the higher it is, the more the plan depends on the peak. The chances are working assumptions for this demo; with historical auction prices they become measured frequencies.

Planning answer

—

Method

Optimization for the plan, price scenarios for the risk

How the plan is computed

The plan comes from linear programming — a technique that, in effect, tries every allowed combination of weekly planting choices and returns the one with the highest revenue. "Allowed" means it respects the physical limits: 42 bays of space, at most 6 kappen planted per week (crew pace), and at least 2 harvested per week (contract commitments). Revenue counts at the harvest week's price. The Women's-Day analysis adds stochastic modeling: the same search is run against the three price futures at once, so the chosen plan is judged not only on its best guess but on how it holds up when prices swing — that is how risk gets a number instead of a feeling.

From revenue to profit

Prices here are estimates and the model currently maximizes revenue. Each new data stream completes more of the risk landscape: realized weekly auction prices turn the three futures into measured likelihoods, and cost and energy data turn revenue into profit — with the footprint line joining when the energy stream lands (see Calibrate).

Model: integer MILP (HiGHS); the uncertainty block solves the mean-CVaR extension of the same model. Validation: the engine reproduces the baseline weekly plan (revenue 2025–2029, €22.24M) before optimizing; the optimized rows change only the stated constraints. Scenario probabilities are working assumptions, not fitted. Basis: figures are totals over the four planting years 2025–2028; harvests of the last plantings land into spring 2029 (≈ €1.35M, included in the totals).