D605 Optimization, catalog number DTAN 6217, is the three CU course in the WGU Master of Science, Data Analytics that applies iterative algorithmic maximization and minimization to a wide class of business problems. Most of the difficulty is not in the solver. It is in the translation step: turning a messy operational situation into an objective function and a set of constraints that an evaluator agrees actually represent the business.
Formulation is where the marks are
Optimization work has an unusual grading profile. Once a problem is correctly formulated, the solving is often a handful of lines. Everything that can go badly wrong happens earlier, in the modelling decisions, and those decisions are made in prose before they are made in code.
The scored questions are these. What exactly is being maximised or minimised, in units a manager would recognise? Which quantities are decision variables you control and which are parameters you merely observe? What limits are hard constraints that cannot be violated and which are preferences that belong in the objective? And what did you leave out of the model, knowingly?
That last one carries more weight than students expect. Every optimization model is a simplification, and the difference between a graduate submission and an undergraduate one is often a single honest paragraph naming what the formulation ignores and why ignoring it is defensible for this decision.
Turning scored aspects into a section plan
Scoring detail sits in your Course of Study rather than the public catalog. Open it, count the scored aspects, and use them as headings worded close to the rubric's own nouns. Each aspect is judged independently against a three point scale and each needs a 2 for the task to pass, so an elegant solution with an unexplained constraint set still comes back.
Optimization submissions carry a specific mapping hazard: the formulation lives in mathematics, the implementation lives in code, and the interpretation lives in business language. An aspect can be satisfied in one of those three registers and invisible in the other two. Restate each formulation decision in plain sentences beside the notation.
The word budget, worked. Take six scored aspects and about 1,900 words of narrative alongside the model. Reserve 150 words to state the business decision and 120 to close on the recommendation, leaving roughly 1,630, near 270 words per aspect. Then shift 60 words out of each descriptive aspect into the formulation aspect and the sensitivity aspect, taking each to about 360. Formulation at 120 words is a restatement of the prompt. Sensitivity analysis omitted entirely is the most common single reason an optimization task is returned.
A structure that fits an optimization deliverable
Where the directions give a structure, use it. Where they do not, this ordering matches how optimization aspects tend to be written and read.
| Section | What belongs in it | How it gets read |
|---|---|---|
| Business problem | The decision, who makes it, how often, and what better means to them | Defines whether your objective function is the right one |
| Decision variables | What you actually control, with units and permissible ranges | Confusing a parameter with a variable invalidates the whole model |
| Objective function | The quantity being maximised or minimised, stated in notation and in words | Scored on whether it matches the business definition of better |
| Constraints | Each limit with its source, its type and what happens if it binds | Constraints with no stated source read as invented |
| Method and implementation | Solver or algorithm, why it suits this problem class, convergence evidence | Checked for fit between problem type and technique |
| Results | The optimal solution, the objective value, which constraints bind | Binding constraints are the most useful managerial output and are often omitted |
| Sensitivity and recommendation | How the solution moves when parameters shift, and what you advise | A point solution with no sensitivity is fragile advice |
State the model twice: once in notation for precision, once in sentences for a reader who will not read notation. Optimization reports that only exist in symbols routinely lose the interpretation aspects.
Evidence craft for a formulated model
The evidence in an optimization task is unusual because much of it is assumption. Assumptions are legitimate evidence when they are declared.
- Give every parameter a source. A cost, a capacity or a demand figure that appears without provenance undermines the whole solution.
- Declare every assumption as its own numbered statement, and say what would change if it were false.
- Report convergence rather than assuming it. Iteration counts, tolerance and any solver warning belong in the narrative.
- Show the units and check them. Unit mismatches between an objective and a constraint are common and produce plausible looking nonsense.
- Cite the method to a source in operations research rather than to solver documentation.
- Use APA for external sources, and keep quoted definitions short.
The most valuable sentence in an optimization report names the boundary of validity: the range of parameters over which your recommendation holds. It converts an answer into advice, which is what a graduate analytics course is asking for.
What separates Competent from a submission sent back
Aspects score independently, so returns here are typically one section deep.
- The objective function is defended against at least one alternative definition of better.
- Every constraint traces to a stated business or physical limit rather than to convenience.
- The results section identifies which constraints bind and what that means operationally.
- Sensitivity analysis exists and produces a statement about robustness, not just a table.
- The recommendation is specific enough to be refused, with an owner and a trigger for revisiting it.
Performance assessment work can be revised and resubmitted with no grade penalty, which makes a return a delay rather than a mark against you. With six month flat rate terms, the cost is entirely calendar, and calendar is what decides how many courses close.
Six mistakes that cost time in D605
- Optimising the wrong quantity. Minimising cost when the business optimises service level produces a technically correct model that answers nobody's question.
- Constraints with no source. A limit invented to make the model tractable has to be declared as such.
- Skipping sensitivity analysis. This is the most frequently missing section in the course and among the easiest to supply.
- Reporting only the objective value. Managers act on which constraints bind and what a unit of relaxation is worth.
- Mixing units. Hours against dollars against units in one expression is a silent error that survives all the way to a wrong recommendation.
- Presenting a point solution as certainty. Parameters are estimates. A recommendation that ignores that reads as naive.
Translating a business situation into a model
The step that decides a D605 grade happens before any notation is written, and it is worth doing on paper. Take the situation as described and answer five questions in order.
First, what decision is being made, and by whom? A scheduling manager choosing shift patterns and a finance director choosing capital allocation produce entirely different models from the same operational description. Second, what does better mean to that person? Cost, throughput, service level, risk and utilisation are all plausible objectives, and choosing the wrong one produces a model that is internally flawless and practically useless.
Third, what can they actually change? This is the decision variable test, and it is where most formulation errors originate. Demand is usually observed, not chosen. Staffing levels usually are chosen. Machine capacity may be either, depending on whether the time horizon allows purchase. Getting this wrong turns a parameter into a lever the business does not have.
Fourth, what limits are genuinely hard? A legal maximum on consecutive hours is hard. A preference for weekends off is soft and belongs in the objective with a weight, not in the constraint set. Modelling a preference as a hard constraint is the classic route to an infeasible model, and infeasibility discovered at the end of a build costs days.
Fifth, what are you deliberately excluding? Every model ignores something: seasonality, employee skill differences, transport times, the fact that people call in sick. Listing those exclusions with a one line reason each is a section that takes twenty minutes and reliably earns credit, because it demonstrates that the simplification was a decision rather than an oversight.
Answer those five in plain sentences, get them checked, and only then write the formulation. Notation produced from a settled understanding is quick. Notation produced while still deciding what the problem is tends to be rewritten three times.
How support works on this course
Send the rubric from your Course of Study and the task directions. What comes back is aspect mapped: a formulation written in both notation and plain sentences, constraints traced to sources, results reported with binding constraints named, a sensitivity section that concludes something, and a walkthrough so the model is yours to defend.
D605 rewards a slow start. An hour spent arguing about what better means for the business saves days of solving the wrong problem, and in a six month flat rate term those days are the entire currency.
Questions students ask about D605
Is D605 the same course as DTAN 6217?
How much mathematics does D605 assume?
Can you solve the model and send me the answer to submit?
Where D605 sits in WGU's programs
The July 2026 catalog places this code in 1 current WGU program. Open a program page for the complete standard path and term positions. The live Degree Plan remains authoritative after transfer credit, substitutions, and mentor planning.
The assessments, one by one
The public catalog does not publish this course's PA/OA identity or task count. WGU Tutors publishes at most one PA manual per course and only from a WGU-controlled public rubric. Until that source exists, PA help begins from the student's real Course of Study and OA support remains preparation only.