Age-Structured Projection Model¶
Model — instantiates Cohort-Structured Replenishment Stabilization
Projects a replenished stock forward one age class at a time, so today's cohort sizes surface as tomorrow's abundance or gap instead of hiding inside a single healthy-looking total.
A total tells you what a stock is today; it says nothing about the shape of what is coming. Age-Structured Projection Model is the calibrated state model that fixes this: it holds the stock as a distribution across age classes, and it steps that distribution forward — each class ageing into the next at some survival/progression rate, a new cohort entering at the bottom — so that the consequences of a thin or fat class become visible years before they arrive. Its defining move is that the future is a function of the age structure, not the aggregate: two stocks with identical totals but different age profiles project to entirely different futures. Where a scenario simulation runs this model under many what-ifs, the model itself is the underlying bookkeeping — the transition structure and the small, bounded places where the current total is allowed to feed back on the rates.
Example¶
A state education agency is told its teacher workforce is "fully staffed" — the headcount matches budgeted positions. The Age-Structured Projection Model represents the workforce as teachers by age band, applies each band's observed retirement and attrition rates, and feeds in the current hiring rate at the bottom. Stepped forward a decade, the picture the total concealed appears: a wave of retirements from a large 55–60 band, and behind it a hollow mid-career band left by a hiring freeze fifteen years ago that no current total reflects. The projection is built cohort-component style[1] — the same age-class bookkeeping demographers formalize as a Leslie matrix — so the freeze-era gap is carried faithfully as those teachers age rather than being averaged away.
The output is a projected age structure, year by year: not a verdict, but the substrate every downstream decision needs. It is what tells the agency the staffing problem is not how many but which classes are missing and when the gap lands.
How it works¶
- Carry the state as an age vector. The stock is a distribution over age/stage classes, not a scalar; the whole method exists to keep that structure intact through time.
- Step with a transition structure. Each class moves to the next by a survival/progression rate; a fresh cohort enters at the base. Diagonals are cohorts ageing — the model's memory of past entry windows.
- Bound the current-stock feedback. Where the current total genuinely feeds back on the rates (density dependence, congestion), that coupling is encoded and explicitly bounded, so dynamics stay cohort-driven rather than total-driven.
- Produce a trajectory, not a forecast fan. The model yields one calibrated baseline path; turning it into a probabilistic forecast is a separate step performed by running it.
Tuning parameters¶
- Age-class granularity — annual classes track sharp cohorts but demand more data; broad bands smooth over the very structure the model exists to show.
- Projection horizon — short horizons are well-constrained; long ones expose the accumulated error of assuming today's rates hold.
- Feedback strength — how much current-stock density-dependence is allowed into the rates; turn it up and the model self-corrects, turn it down and cohort echoes dominate.
- Entry assumption — how the incoming cohort is set (fixed rate, replacement, policy input); the single most consequential dial for the far end of the horizon.
- Calibration window — how many past years the rates are fit on; a short window tracks regime change, a long one is steadier but stale.
When it helps, and when it misleads¶
Its strength is making the age structure load-bearing: a retirement cliff or a hollow class that a total averages into invisibility becomes an explicit, dated feature the rest of the archetype can act on. It is the shared substrate the strength register, the decision rules, and the scenario runner all lean on.
Its central failure mode is treating a single projected line as a forecast. Rates are assumed roughly stationary, and the moment the regime shifts — a policy change, a demand shock — the clean trajectory misleads with false confidence; the outputs are only as good as the calibration window and the entry assumption. The classic misuse is to freeze the rates, read the baseline as prophecy, and skip the uncertainty. The discipline is to treat the model as a structure whose assumptions must be stated, and to hand the range of futures to the scenario runner rather than pretending the baseline is one.
How it implements the components¶
Age-Structured Projection Model realizes the state-and-dynamics core of the archetype — the parts a calibrated model can carry, not the whole machinery:
age_structure_memory_model— the transition structure is this: it remembers which cohorts entered when, and carries each forward as it ages, so past entry windows persist as diagonals in the state.current_stock_feedback_boundary— it encodes where the current total feeds back on the rates and explicitly bounds that coupling, keeping the dynamics cohort-driven rather than governed by the aggregate.
It does not implement the forward echo forecast under uncertainty — that is Cohort-Echo Scenario Simulation, which consumes this model — nor cohort measurement (Cohort Strength Table) or any policy response.
Related¶
- Instantiates: Cohort-Structured Replenishment Stabilization — this model is the state engine the appraisal projects the stock through.
- Sibling mechanisms: Cohort-Echo Scenario Simulation · Year-Class or Vintage Matrix · Cohort Strength Table · Early-Window Sentinel Monitoring · Age-Weighted Quota or Capacity Rule · Weak-Cohort Trigger Rule · Strong-Cohort Pacing Rule · Cohort-Diversified Source Plan · Recruitment-Failure Postmortem
Editorial Notes¶
Form Classification¶
Form family: Analysis, Modeling & Optimization
Rationale: The mechanism projects a replenished stock forward one age class at a time, so today's cohort sizes surface as tomorrow's abundance or gap instead of hiding inside a single healthy-looking total, so its operative form is offline analysis, modeling, or optimization.
Independent corroboration: The frozen evidence defines Age-Structured Projection Model as 'Projects a replenished stock forward one age class at a time, so today's cohort sizes surface as tomorrow's abundance or gap instead of hiding inside a single healthy-looking total', so its operative form is Analysis, Modeling & Optimization.
Review outcome: Independent reviewer agreement; high confidence.
Origin Attribution¶
Primary origin: Biology & Ecology
Origin pattern: Convergent development
Present-day reach: Multi-domain
Rationale: Population ecology projects abundance by age-specific survival and fecundity, canonically using Leslie matrices over a cohort vector.
Related originating lineages:
- Marine Science & Oceanography — Age-structured fishery stock assessment is a mature implementation.
- Mathematics — Matrix population models and eigenanalysis formalize transitions, stable age distributions, growth rates, and sensitivity.
- Sociology & Anthropology — Human demography's cohort-component method projects age-structured populations, migration, and dependency composition.
- Statistics & Experimental Design — Estimation of age-specific transition rates and uncertainty contributes empirical calibration.
Review resolution: Population ecology and human demography converged on cohort-structured projection, with matrix mathematics and statistical estimation as formative foundations and marine stock assessment as a mature independent development. Agricultural and climate uses are applications rather than separate origins.
Attribution caveat: The Leslie-matrix formalism is population-biological, with a closely parallel demographic lineage.
Review outcome: Reconciled after independent review; high confidence.
Notes¶
The model is an object, not a forecast: it defines how the stock moves, but it produces a single calibrated trajectory. The range of plausible futures — and the tail risks that matter for buffering — come from running it, which is Cohort-Echo Scenario Simulation. Keeping the two separate lets the transition structure be improved (better rates, finer classes) without re-arguing every scenario.
References¶
[1] Leslie, P. H. "On the Use of Matrices in Certain Population Mathematics". Biometrika 33(3), 183–212 (1945). Formalizes age-structured population projection with a matrix that advances the retained age distribution rather than collapsing it into an average. registry ↩