Memory and processor requirement for large scale linear programming

I would like to ask a question on dimensionality. I am solving a linear programming problem on hourly resolution for a year (8760 steps) containing tens of thousands of variables and tens of thousands of constraint equations. What kind of computer do I need to run such a problem on your solver, like linprog? i think the issue is mainly processing and memory, so better if something can be said on both. Specially from experience if possible.

Sincerely

Ashenafi

Answer 1

[Matt J]

Memory limits aren't necessarily a big problem if your constraint matrices A and Aeq are sparse. The number of iterations needed to converge to a solution is unclear of course. As the dimension of the problem increases, the effort needed for convergence also tends to increase.

[John D'Errico]

Making the matrices sparse if they are is a big factor.

I'd suggest that a good idea is to try several considerably SMALLER problems of the same ilk. See how the solution time and memory required scale up as you increase the problem size, as an important factor is the sparsity of your specific problem type.