Act as an operations research analyst helping me frame and solve a decision problem. Do not pretend to have run an optimization model, accessed data, or found an optimal solution unless the calculations and inputs are shown. DECISION TO MAKE: [DECISION] OPERATING CONTEXT: [OPERATING CONTEXT] DATA AND UNITS: [DATA] CONSTRAINTS AND BUSINESS RULES: [CONSTRAINTS] OBJECTIVE AND SUCCESS METRIC: [OBJECTIVE] IMPLEMENTATION WINDOW: [IMPLEMENTATION WINDOW] If a required input is missing, ask up to 3 clarifying questions. Otherwise, produce an analyst-ready decision memo with these sections: 1. Decision statement: define the choice, decision owner, scope, time horizon, and what is explicitly out of scope. 2. Model specification: identify decision variables, objective function, constraints, assumptions, units, and data gaps. Use equations or pseudocode where helpful. Classify the problem as linear/integer programming, network flow, queueing, simulation, forecasting, or another appropriate method, and explain why. 3. Baseline and scenarios: describe a no-change baseline plus three feasible scenarios. State what must be calculated or simulated for each, including cost, service level, utilization, risk, and sensitivity to uncertain inputs. 4. Recommendation: give a conditional recommendation only to the extent supported by the supplied data; otherwise state the next analysis needed. Include owners, ordered steps, dependencies, and dates within [IMPLEMENTATION WINDOW]. 5. Validation plan: specify back-testing, pilot design, success thresholds, monitoring cadence, and rollback triggers. Avoid false precision. Keep dollars, hours, units, and percentages explicit. Flag constraints that are mutually inconsistent, objectives that conflict, and variables that cannot be controlled. Before answering, check dimensional consistency in every formula and confirm that the recommendation does not exceed any stated capacity, budget, service, or policy constraint.
Fill in
| Placeholder | What to enter | Example |
|---|---|---|
| [DECISION] | State the specific operational choice that must be made. | Choose daily delivery routes and driver shifts for 42 grocery deliveries across Brooklyn and Queens. |
| [OPERATING CONTEXT] | Describe the process, locations, teams, customers, and current operating problem. | A local grocer has six drivers, one dispatch lead, 9 a.m.–7 p.m. delivery windows, and frequent late afternoon routes. |
| [DATA] | Paste available volumes, costs, capacities, demand patterns, service levels, and their units. | Average stop time 12 minutes; travel-time matrix by neighborhood; driver cost $26/hour; 42–58 orders/day; late-delivery rate 14%. |
| [CONSTRAINTS] | List hard limits and business rules such as staffing, delivery windows, budget, compliance, and contracts. | Each van holds 20 totes; drivers work no more than 8 hours; cold orders cannot be in a van more than 90 minutes; budget increase capped at $600/day. |
| [OBJECTIVE] | Define the primary goal and how success will be measured. | Reduce late deliveries below 5% while minimizing daily labor and mileage cost. |
| [IMPLEMENTATION WINDOW] | Give the deadline, pilot duration, and any decision dates. | Recommend a pilot by October 14; run it for two weeks; decide on rollout by October 31. |
How to use
- Paste the real decision and retain units beside every number, including the time period it covers.
- Check the model section against constraints owned by finance, operations, labor, and customer-service teams.
- Ask the decision owner to validate assumptions before treating any scenario as a recommendation.
- Follow up with: “Create a spreadsheet-ready input table and formulas for the baseline and three scenarios.”
Variations
Staffing model
Use this to size shifts against demand and service targets.
Build a staffing-analysis plan for this operation. DEMAND BY INTERVAL: [DEMAND] CURRENT STAFFING: [STAFFING] SERVICE TARGET: [SERVICE TARGET] LABOR RULES: [LABOR RULES] Return a queueing or capacity model specification, a 30-minute staffing table, three staffing scenarios, and a pilot plan. Show workload assumptions, shrinkage, breaks, training time, and occupancy targets. Do not claim a forecast is accurate without validation data. Ask up to 3 questions only if required input is missing. Check each recommended shift against labor rules and total paid hours.
Inventory policy
Use this to set reorder rules for items with uncertain demand or lead time.
Design an inventory-policy analysis for these items. ITEM DATA: [ITEM DATA] DEMAND HISTORY: [DEMAND HISTORY] LEAD TIMES: [LEAD TIMES] SERVICE TARGETS: [SERVICE TARGETS] CONSTRAINTS: [CONSTRAINTS] Return recommended inputs for safety stock, reorder point, order quantity, review cadence, and exception rules; show formulas and assumptions. Include stockout, obsolescence, and supplier-risk scenarios. Ask up to 3 questions only if required input is missing. Check that demand and lead-time units match and flag data too sparse for a reliable parameter estimate.
Executive decision memo
Use this after a model exists and leaders need a short decision document.
Turn this completed operations analysis into a decision memo for leadership. ANALYSIS: [ANALYSIS] DECISION OWNER: [OWNER] OPTIONS: [OPTIONS] RISKS: [RISKS] DEADLINE: [DEADLINE] Produce a one-page memo with recommendation, baseline, options comparison, economic and service impact, assumptions, risks, implementation owners, and the exact approval requested. Keep model detail in a short appendix. Do not add savings or performance claims not present in the analysis. Ask up to 3 questions only if required input is missing. Check that the requested decision is actionable by the named owner and that all numerical claims have a stated source.
Tips
- Write hard constraints separately from preferences; a solver can trade off a preference but must never violate a true limit.
- Normalize every input to compatible units before modeling, such as orders per 30 minutes and labor hours per shift.
- Ask for a baseline first, because a proposed optimization is meaningless without comparison to the current rule.
- Pilot against predefined service, cost, and safety thresholds before changing a live operating process.
FAQ
Can ChatGPT solve an optimization problem directly?
It can formulate a small, transparent model and help create code or spreadsheet logic. For material decisions, run the model in a reproducible tool and validate the data and constraints.
What data do I need for operations research?
At minimum, define demand, capacity, costs, time horizon, constraints, and the objective. Historical variability and current performance data make recommendations more credible.
Should I optimize cost or service level?
Usually set one as the objective and the other as a constraint or service threshold. If both matter, state the tradeoff explicitly and compare a small set of feasible scenarios.