Pyomo - MINLP

[CHITHRAPONNU R 20PHD1199]

Is it possible to solve MINLP using Pyomo?

Can anyone help me solve my problem? 

Disclaimer:
This message was sent from Vellore Institute of Technology.  The contents of this email may contain legally protected confidential or privileged information of “Vellore Institute of Technology”.  If you are not the intended recipient, you should not disseminate, distribute or copy this e-mail. Please notify the sender immediately and destroy all copies of this message and any attachments. If you have received this email in error, please promptly notify the sender by reply email and delete the original email and any backup copies without reading them.

[Venu©]

Pyomo is just programming way based on python. To solve MINLP problem u have to use solver and mostly all solvers can be used through pyomo

--
You received this message because you are subscribed to the Google Groups "Pyomo Forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pyomo-forum...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/pyomo-forum/8743fb1b-9175-473d-b719-480e63d8142en%40googlegroups.com.

[CHITHRAPONNU R 20PHD1199]

Yes. Thank you for your response, firstly. Yes, I used IPOPT to solve MINLP. But it's not working. Can you please suggest any other with a simple example? 

To view this discussion visit https://groups.google.com/d/msgid/pyomo-forum/CABOznq_7sP9JNCspuWf8hiJYs27C6MFqOebCqHQkDCFpeV%3Deew%40mail.gmail.com.

[Venu©]

So I believe some of ur decision variables must be binary and ipopt will not give any output with binary options. So u can use different solvers like Baron or Gurobi to solve MINLP problem. There r some open source solvers available for solving MINLP problem. U can check AMPL website to get list of open source solvers and AMPL does offer some. 

To view this discussion visit https://groups.google.com/d/msgid/pyomo-forum/CALxzMWciD%2BgHhM7ShcgaO811Hw7-ni%2BFeBfDR_BK3hL9vuEK3w%40mail.gmail.com.

[Ahmad Heidari]

CHITHRAPONNU,

Since non-linear programming usually provides a local optimum, other linearized methods are applied. 

You can apply some methods to solve MINLP problems, though. 

First, you might need to linearize the nonlinear terms using well-known mathematical methods.

You can use some advanced optimization techniques, such as Benders Decomposition, to increase the probability of reaching a global optimum.

In case of solvers, as @Venu correctly mentioned, you can sign up on the AMPL website and obtain a 30-day free trial of commercial solvers. 

In your case, for MINLP problems, the Knitro solver works best. You may use Gurobi with some parameter tuning, as well. 

[CHITHRAPONNU R 20PHD1199]

Thank you @venu and @heidari. Conversion of my model nonlinear to linear is little complicated. I'll first try @venu suggestions. If it's not working for me I'll try @heidari suggestions. 

To view this discussion visit https://groups.google.com/d/msgid/pyomo-forum/24b3d764-fbe7-4a8a-9407-c10e976fcc06n%40googlegroups.com.