search

BYU-PRISM / GEKKO

GEKKO Python for Machine Learning and Dynamic Optimization
https://machinelearning.byu.edu
Other
594 stars 104 forks source link

Parallel multiprocess on multiple cores #161

Open farhadbl opened 1 year ago

farhadbl commented 1 year ago

Not a bug, maybe. a feature to add or (maybe already developed) to let the users to run on multiple cores or run it on servers to speed up the optimization convergence.

APMonitor commented 1 year ago

Thanks for that feature request. Most of the time is typically spent with the solver. There are parallel options for solvers such as IPOPT to use ma77, ma97, etc. We found that these linear solvers so speed up the solution, but not by more than a factor or 2-3 for the solver. Here are options for parallel solution with Gekko if you are solving multiple separate optimization problems.