Open source Solver to replace CPLEX in example?

Specific Domains Optimization (Mathematical)
,
1

Ten years ago this week, I read an article entitled Optimization in Julia. It was my first introduction to Julia and a great example of how easy one can define a constrained optimization problem in JuMP. The example used IBM’s commercial solver CPLEX (free for .edu), IIRC the only solver at the time that could find a solution.

I’m wondering if today there are any open source solvers that could replace CPLEX in this example. Suggestions?

2

I would 100% recommend HiGHS via HiGHS.jl

HiGHS is pretty the best open-source LP solver at the moment, and it also supports MILP and QP problems.

3

Hi @rt9f, welcome to the forum :smile:

Yes, definitely use HiGHS. The example you linked is a nice example I hadn’t seen before. I’ll write it up as a tutorial for the JuMP documentation (with a few syntax changes that have happened since 2016).

4

So I wrote this up [docs] add sports scheduling tutorial · Pull Request #4160 · jump-dev/JuMP.jl

It’s actually a really tricky problem (as formulated) to solve to optimality.

5

Hi @odow,
Nice done!
Thank you for quick reply.

6

Here’s the link for posterity

7

Thanks. Few minor points:

  • there is a couple #!format:off, / on left
  • for international public a rewrite/ explanation of the Big10 /Conference/ “back-to back away games” would make the post more accessible…
  • perhaps a note that this is a linear problem with binary variables so we can use Highs that is a linear solver capable of representing discrete variables (or something on the line)