Weekly Take-home

The fourteen regular software take-homes, worth 55 points.

One Excel workbook a week, three sheets: Instructions (every step and why it matters), the model sheet where you do the work, and Free responses for your written answers plus the required AI-use disclosure (see Using AI).

The model is not built for you. Every model sheet works in the same four blocks:

  • A. Case brief: the narrative and raw figures, exactly as somebody in the case would hand them to you. Not every figure is one you need, and at least one arrives on the wrong basis or in the wrong unit.
  • B. Your model: before computing anything, define the quantities in words with their units, write the relationships or constraints you will use, and say which figures you left out and why. Most of the thinking is here.
  • C. Solve: the work area. Nothing is pre-labelled with the answer’s shape.
  • D. Read the answer: the few numbers a decision-maker would actually act on.

Fill every red cell; plain cells are given. Submit the completed .xlsx on Canvas by Sunday at 11:59 PM of that week. That one file is your entire submission, except in Weeks 5 and 6, which also ask for a picture of your decision tree from the tree builder. Fourteen weeks carry one, Weeks 1–13 and 15, and they are half the question pool for the two exams. Points are in Grading; late work follows the 3-3-3 rule, where the trap worth knowing is that a placeholder upload becomes your graded work.

Each row links to that week’s take-home; the starter file downloads from the lesson page.

Wk Weekly take-home (graded)
1 Break-even from a brief: classify the costs, chart it, then re-run it on what really happened
2 Probability from raw counts, put on the decision's own time basis, then EV, spread, and a price ceiling
3 Build the payoff table, work all five criteria, then find where Hurwicz changes its mind
4 Build the cost table: EV, EOL and the EVPI ceiling
5 Sequential tree: fold back, price the option, find the flip and the EVPI
6 Bayesian revision from counts, EVSI, and a utility re-fold
7 Formulate two LPs and enumerate every corner (no Solver)
8 Formulate an LP, wire up Solver, read the Answer Report
9 LP sensitivity: shadow prices, both ranges & reduced cost
10 Three network models with Solver: an unbalanced grid, matching & a hub
11 Weighted scoring, AHP consistency & preemptive goal programming
12 CPM / PERT from a narrative: critical path, risk & crashing
13 Choose the window and alpha: moving average vs smoothing
15 Linear trend and a negative-slope regression: R² & forecasts

Week 14 is Thanksgiving break and Week 16 is review, so neither has a take-home.