Free Math Solvers for Operations Research

On harumi.io, users can model and solve optimization problems directly in our platform using both commercial and free solvers. Among the free options, three stand out for their reliability and accessibility: GLPK, CBC, and HiGHS. They allow companies and researchers to start experimenting with Operations Research (OR) without additional licensing costs.

GLPK (GNU Linear Programming Kit)

Scope: Linear Programming (LP) and Mixed Integer Programming (MIP).

Strengths: A stable and widely adopted solver, especially in academic and research environments. Works seamlessly with modeling frameworks like Pyomo, making it ideal for teaching and prototyping.

Considerations: Performance can lag for very large or complex MIPs, but it’s robust for small to medium problems.

CBC (COIN-OR Branch and Cut)

Scope: Mixed Integer Linear Programming (MILP).

Strengths: Part of the COIN-OR open-source project, CBC is often the default solver in Python OR packages. It offers a solid mix of usability and performance.

Considerations: Best suited for medium-sized models; runtime grows with highly combinatorial problems.

HiGHS

Scope: LP, MIP, and quadratic programming.

Strengths: A modern solver designed for speed and scalability. Recently integrated into SciPy, HiGHS bridges the gap between the OR community and data science.

Considerations: Although newer than GLPK and CBC, it is advancing quickly and already delivers competitive results.

With GLPK, CBC, and HiGHS available directly on harumi.io, anyone can explore Operations Research models at no cost. They are excellent starting points for learning, prototyping, and in many cases, solving real business problems without the overhead of commercial licensing.