search

BYU-PRISM / GEKKO

GEKKO Python for Machine Learning and Dynamic Optimization
https://machinelearning.byu.edu
Other
573 stars 102 forks source link

APOPT IPOPT output #101

Open strahl21 opened 3 years ago

strahl21 commented 3 years ago

Not a bug in the code. I noticed the following output when I was trying to use IPOPT in estimation mode for the temperature control lab (Lab C) with remote=False (so it defaults to APOPT): APOPT_IPOPT_info

It could be confusing for some because Gekko reports that it does not support IPOPT and instead uses APOPT, but then in the output from the solver, it lists IPOPT as the solver (although clearly from the output format APOPT was actually used). The two should be consistent.

The code used was the code under "Lab C: Python GEKKO Moving Horizon Estimation". Lab C

loganbeal commented 3 years ago

@strahl21 Which OS are you using?

strahl21 commented 3 years ago

Ubuntu 18.04

APMonitor commented 3 years ago

With remote=True (using the public server), there is the option to use IPOPT. For local solve, I have to distribute a separate version of the local executable that includes IPOPT because the libraries are really big and I can only distribute it freely with the MUMPS version. Windows includes IPOPT because I didn't need to distribute many of the compiler libraries with IPOPT and it only adds about 3 MB to the executable size. Thanks for recommending that the solver is reported correctly. I reference the global SOLVER option that I didn't want to change because it is input only. I'll add this to the list of things to fix.