This page has five Gemini prompts for Excel Solver. They cover the whole setup: framing the objective, writing the constraints, choosing a solving method, reading the answer report, and moving the model to Python if you outgrow the add-in. Copy a prompt, describe your numbers, and paste it into Gemini.
What to know before you paste
- Solver ships as an Excel add-in, so tell Gemini your Excel version. The menu names and the enable steps change between versions.
- Describe your decision variables and what “best” means before asking for the setup. Solver is only as good as the problem you frame.
- If you plan to move the model to a script, say so early. The prompt can then shape the problem for Python or R instead of the Solver dialog.
The prompts
1. Turn a goal into a Solver setup
Help me set up Excel Solver. My goal is to [objective, e.g. maximize profit]. The cells I can change are [decision variable range], and the cell that measures the goal is [objective cell]. Describe the objective cell setting, the variable cells, and any bounds I should add, in the order I enter them in the Solver dialog.What it does: Converts a plain goal into the exact fields Solver asks for.
How to use: replace [objective], [decision variable range], and [objective cell] with your real ranges.
2. Write the constraint list
Here is my problem: [problem summary]. The limits are: [list limits, e.g. budget, capacity, hours]. Turn each limit into a Solver constraint with the correct cell range, operator, and value. Flag any limit that could make the problem impossible to solve.What it does: Turns messy business limits into clean constraints and warns you about conflicts.
How to use: replace [problem summary] and [list limits]; add units so the model keeps quantities consistent.
3. Choose the right solving method
My model is [linear or nonlinear] because [why]. Describe the differences between Simplex LP, GRG Nonlinear, and Evolutionary, and tell me which one to pick for this model and why. Also note which settings to change, like precision or convergence, for [my tolerance need].What it does: Picks the method that will actually converge instead of running forever.
How to use: replace [linear or nonlinear], [why], and [my tolerance need] with your model’s details.
4. Read the Solver answer report
Here is my Solver answer report: [paste report]. Explain each section to me in plain language, tell me whether the solution is optimal or just feasible, and point out the binding constraints. If the status is not optimal, suggest the most likely reason and one thing to change.What it does: Translates a dense report into what you did and did not achieve.
How to use: replace [paste report]; if the report is long, paste the answer report and the constraint table.
5. Move the model to Python
Rewrite this Solver model as Python: objective [objective], variables [variable range], constraints [constraints]. Use [library, e.g. PuLP or SciPy]. Return a short script with the data at the top so I can edit values without touching the logic, and print the optimal result and the active constraints.What it does: Gives you a version you can run on bigger data than the add-in handles.
How to use: replace [objective], [variable range], [constraints], and [library] with your model and chosen package.
FAQ
Do I need the Solver add-in installed to use these prompts?
To solve inside Excel, yes. You enable it once from the add-ins menu. The prompts help you frame the problem, which is the hard part either way.
Why does Solver return “Solver could not find a feasible solution”?
Usually two constraints conflict, or a bound is too tight. Ask the model to relax one constraint at a time and report which change makes the model solvable.
What is the difference between the LP and nonlinear methods?
Simplex LP is fast and exact but assumes straight-line relationships. Nonlinear methods handle curves and ratios but can settle on a local answer, so the starting values matter.
Can I use these prompts for Google Sheets?
Sheets handles smaller models. Name it as the tool, and the prompt will suggest the built-in solver or a manual setup instead of the Excel dialog.
How do I know the answer is actually the best one?
Check the report status, then test a couple of nearby values by hand. If small changes do not improve the objective, you are likely at a true optimum.
Take what works here and paste it straight into your tool of choice.