Gurobi Operations Research Modeling

Problem

Operations research modeling often breaks down when translating a business description into a solvable model: decision variables, objective terms, and constraints are easy to miss, while parameters live scattered across Excel, CSV, and JSON files. After solving, it is still hard to verify status, objective value, variable values, and constraint validity. gurobi-optimization-skill turns this workflow into a constrained engineering pipeline.

How it works

The skill follows problem intake → data validation → model formulation → code generation → solve validation → result analysis, with a hard rule that parameters must come from user-provided files or conversation rather than invented defaults. Key capabilities include:
- Unified data entry: data_loader.py and data_validator.py handle .csv, .xlsx, .json, and .dat, then check required fields before modeling.
- Guided model building: step-by-step prompts/ files drive Gurobi Python code generation and write the model to outputs/{problem}/model.py.
- Solve and validation: after running the model, model_validator.py checks status, objective value, Gap, variables, and constraint results.
- Iteration and solver fallback: when the user asks to adjust constraints, change objectives, or add variables, it uses iteration_context.md plus tuning or revision prompts, saves timestamped new files, and can switch between Gurobi, SCIP, OR-Tools, and CBC via solver_factory.py.

Boundaries

It is best for LP, MILP, QP, and MIQP-style tasks centered on Gurobi, with template support for production scheduling, logistics routing, resource allocation, and inventory management. Without a Gurobi license, avoid relying on gurobipy as the only path; missing business data should be requested first; and placeholder data must be explicitly marked and replaced.