← Past problems

HiMCM 2012 · Problem set

A wildlife-management problem paired with a household-finance forecasting problem. Problem A asks teams to model the reintroduction of elk into Great Smoky Mountains National Park, decide whether the herd will persist or die out under current conditions, and recommend interventions that improve long-run growth. Problem B asks teams to build a simple, defensible rule a consumer can follow each week — buy a full tank or a half tank? — based on observable weekly U.S. gasoline-price data, then validate the rule out-of-sample against a different year and at least one major-city market.

Contest datesNovember 8 – November 19, 2012 (extended weekend window) [illustrative]
Participation~450 teams, predominantly United States and China [illustrative]
Problem AAmerican Elk — stage-structured population model, viability under reintroduction, intervention recommendations
Problem BHow Much Gas Should I Buy This Week? — weekly price series, decision policy (full vs. half tank), out-of-sample validation
Official results 2012 HiMCM problems & commentary
Both 2012 prompts are indexed on COMAP's previous problems page. Read the official statement before our outline — the outline is most useful as a second pass.

The two problems

Why this year is good practice

  • Two very different model archetypes. A is a stage-structured population model (a Leslie / Lefkovitch matrix with stochastic vital rates); B is a rolling-window time-series decision policy. Together they cover the two pillars judges reward — dynamics over time and decisions under uncertainty.
  • Real public data. EIA weekly retail gasoline series and elk reintroduction reports from the National Park Service are both freely available, so every assumption can be grounded in a citation rather than a guess.
  • Two non-technical deliverables. A demands a Wildlife Commissioner letter, B demands a newspaper-suitable letter. Strong papers nail the executive paragraph and one chart that a non-modeler can read in 30 seconds.