- #1
maistral
- 240
- 17
Hi. I seem to have forgotten how to implement equality constraints to barrier NLPs and quadratic NLPs.
Say for example I have this problem:
Max Z = x12 + 2 x22
ST:
x12 + x22 ≤ 1
x1+ x2 ≤ 1
The unconstrained problem (quadratic penalty - correct me if I'm wrong) then becomes
Z = - x12 - 2 x22 - α [x12 + x22 - 1]2 - β[x1+ x2 - 1]2
I'm wondering how will I implement another constraint in equality format; ie.
Max Z = x12 + 2 x22
ST:
x12 + x22 ≤ 1
x1+ x2 ≤ 1
x1- x2 = 0.5 ← how do I add this constraint in penalty, or barrier form?
Say for example I have this problem:
Max Z = x12 + 2 x22
ST:
x12 + x22 ≤ 1
x1+ x2 ≤ 1
The unconstrained problem (quadratic penalty - correct me if I'm wrong) then becomes
Z = - x12 - 2 x22 - α [x12 + x22 - 1]2 - β[x1+ x2 - 1]2
I'm wondering how will I implement another constraint in equality format; ie.
Max Z = x12 + 2 x22
ST:
x12 + x22 ≤ 1
x1+ x2 ≤ 1
x1- x2 = 0.5 ← how do I add this constraint in penalty, or barrier form?