kint vs.
manual planning.
Manual planning with spreadsheets and experience works, and it carries knowledge no model has. As the number of interacting rules grows, nobody can show any more how far a plan is from the best one.
dimension by dimension.
Properties you can check, not benchmark claims. The last column says which approach fits better on that dimension.
| Dimension | kint | manual planning | Fits better |
|---|---|---|---|
| Plan quality | The solver reports the bound and the gap. At 0.0 %, no better plan exists for the model as stated, within the solver's tolerance. | Usually good, with no way to measure how far from the best it is. | kint |
| Re-planning | Change an input and run the model again. | Rework the plan by hand, often under time pressure. | kint |
| Size | Thousands of variables and rules in one model. | Becomes hard to keep consistent as rules and units grow. | kint |
| Consistency | Each run pins the data version, the model version and the solver settings, so the plan can be re-checked against the record. | Depends on who plans and how much time they have. | kint |
| Explanation | States the binding rule and what would change the plan. | The planner explains from experience, which is hard to audit. | kint |
| Knowledge of the operation | Needs that knowledge written down as rules and data. | Deep knowledge built over years, including unwritten rules. | manual planning |
| Setup | Needs read-only data exports and a model that kint's forward deployed engineers review with the domain experts. | Already in place and understood by everyone. | manual planning |
manual planning: the better fit.
The problem is small and stable, or the rules change faster than anyone can write them down, or nobody outside the team asks why.
the decision has to be defended.
Several rules interact, re-planning happens often, and you have to show a board, a regulator or an auditor why this plan and not another.
The best models start from the planners. kint's decision agent writes down the rules they already apply, kint's forward deployed engineers review the model with them, and the planners decide. The plan comes with its bound and gap; the team keeps the judgement.
kint next to the other approaches.
- kint vs. custom developmentyour own optimization team and solver stack
- kint vs. heuristic methodsrules of thumb, metaheuristics and custom search
- kint vs. general AI and MLLLMs and general machine learning
- kint vs. spreadsheet optimizationbuilt-in spreadsheet solvers and solver add-ins
manual planning and kint, on your decision.
Bring a plan your team made today. We put the kint plan next to it, with the binding limit of each, and you judge.
- All comparisons
- kint next to the alternatives
- Results
- Plan, binding limit, sensitivity, bound and gap, record
- Problem types
- the decisions kint solves
- contact
- hello@kint.pro