Open woodwm opened 7 years ago
It is theoretically possible to correspond to IPOPT. I am interested in how much the difference of solver is effective. But because it needs many changes, I don't intend to respond immediately.
I am sure either. But I have seem that IPOPT is extensively used in solving nonlinear optimal control problem, such as GPOPS 2. One of the main change of GPOPS 2 to GPOPS 1 is adding support to IPOPT. The author also compared the performance of IPOPT and SNOPT. In some cases IPOPT is better, some are not.
FWIW I just asked about getting an academic license for SNOPT as a home-use user unaffiliated with any university/government/business and was told they don't offer that, so its something like $6000 for SNOPT. IPOPT is free.
I find that many optimal control solvers use free and open source IPOPT solver for the NLP problem, for example: PSOPT. So I think it would be great to add support for IPOPT either through PyIpopt or Pyomo.
Thanks.