Hi,
Is it possible to accelerate the solution of several similar lp problems by passing the simplex basis using Julia and the Clp.jl package ?
Laurent
Hi,
Is it possible to accelerate the solution of several similar lp problems by passing the simplex basis using Julia and the Clp.jl package ?
Laurent
It is not possible at the moment. I’d strongly encourage you to use HiGHS.jl instead. This will pass the basis between solves automatically if you modify the problem in-place.
Hi Oscar. Thank you for your help !
I do use HiGHS at the moment. I compare the computation time with a C++ code calling Clp and saving simplex basis from call to call. In some cases the C++ code is faster and I was wondering if the different solvers could be the reason.
Not a big deal anyway. Thank you again.
In some cases it is possible that Clp is faster. I know that for small LPs, HiGHS can be slower than competitors. It has a lot of “heavy” machinery that is useful for large problems but just slows down small problems. But the point is: HiGHS is more actively developed, and likely more numerically robust. Is it worth the performance difference?
Probably not ![]()