Skip to main content

How can I formulate multi-layer optimization?

  • April 10, 2023
  • 4 replies
  • 50 views

Forum|alt.badge.img+4

Here is the problem:

“k” is a parameter in optimization problem A, for a specific “k1”, we can get the optimum solution “S1” for problem “A”. Now I hope the solution “S” is closer to a specific solution “S*”, so I will need to change the “k” to minimize |S-S*|. How can I formulate this double-layer problem in AIMMS?

4 replies

Marcel Hunting
AIMMSian
Forum|alt.badge.img+4
  • AIMMSian
  • 258 replies
  • April 13, 2023

Hi @Chunyang. Without knowing anything about parameter k the best you can do I think is to loop over k, using some fixed interval, and solve the model for each k.


Forum|alt.badge.img+4
  • Author
  • Enthusiast
  • 19 replies
  • April 13, 2023

Hi @Chunyang. Without knowing anything about parameter k the best you can do I think is to loop over k, using some fixed interval, and solve the model for each k.

Yeah, it does a solution, but it means I can’t get the precise value for the best k. Thanks a lot anyway!


Marcel Hunting
AIMMSian
Forum|alt.badge.img+4
  • AIMMSian
  • 258 replies
  • April 13, 2023

Whether k can be used as a variable depends on how k is used in your model. I can only advise on that if you share your model.


Forum|alt.badge.img+4
  • Author
  • Enthusiast
  • 19 replies
  • April 14, 2023

Whether k can be used as a variable depends on how k is used in your model. I can only advise on that if you share your model.

Thanks a lot, I don’t have the model now, it’s just an initial thought. I will try to complete the thought first.



Didn't find what you were looking for? Try searching on our documentation pages:

AIMMS Developer & PRO | AIMMS How-To | AIMMS SC Navigator