KardashevJoin Kardashev Ascent ↗Ascent experiments / 004
A place to store it. The capacity to renew it.
Give a fictional genebank 20 funding units. Choose how much goes to cold storage, laboratory reviews and growing space. Follow the workload that each allocation creates.
A fixed allowance / several demands
Each fictional collection occupies one slot. Its chosen renewal schedule needs a lab review followed by a growing job. More storage can admit more collections, while creating more work for the same lab and fields.
Run the model to see the comparison.
All rows have the same 20-unit allowance and initial candidate list. Your allocation may leave units unspent. “On schedule” includes only admitted collections; the separate “not admitted” count keeps that boundary visible. Overdue collection-years adds the overdue count across all years.
A backlog has a history
On schedule Overdue Not admitted
Try aligning all first deadlines. The same average annual capacity can perform differently when work arrives together. Reviews can also expire while a collection waits for growing space.
Method / genebank workload v1
Admit the smaller of candidate count and storage slots, in ascending collection-ID order. That admitted set stays fixed. Initial deadlines are spread cyclically over years 1–6, or all set to year 1. These are equal fictional workloads, with no crop-specific values or priorities.
At each year's lab phase, review collections due this year, overdue, or due next year, if their previous review is no longer current. A review is current in its completion year and the following year. Choose earliest deadline first, then lowest ID.
In the growing phase, renew due or overdue collections that have a current review, using the same order. Same-year review and renewal are allowed in this annual abstraction.
A renewal at the end of year y sets the next deadline to y + 6 and clears its review record. Any remaining deadline at or before y is overdue. Late collections stay in the queue; the model makes no claim that they are dead or unusable. A fresh review after an earlier one expired counts as a repeated review.
The six-year cycle and two-year review freshness are chosen service rules, unrelated to FAO thresholds or any species' lifespan. There are no germination observations, seed quantities, field failures, pollination constraints, storage outages or genetic-change predictions. Real regeneration should follow biological evidence; maximizing its frequency is not a conservation goal.
Funding units represent a fictional repeated allowance. The conversion rates are teaching choices, not prices, electricity estimates or a real genebank staffing plan. The scheduler is a transparent fixed rule; it is not a trained AI.
Read the model and page code · JSON downloads include inputs, every year's jobs, review evidence and final collection records.
Real conservation / separate evidence
FAO's 2014 Genebank Standards describe active viability monitoring and careful regeneration, including risks to genetic integrity. NordGen explains why depositing genebanks keep testing their own collections and decide when fresh seeds are needed for Svalbard. Those sources inspire the question; they do not supply this model's numerical rules.