Posted by thomas floeck on October 14, 1997 at 04:15:30:
In Reply to: crucial constrained optimization problem posted by Thomas Floeck on October 13, 1997 at 06:37:32:
: Hi networld,
: is there anyone familiar with solving following problem:
: f=min s.t. g>0 by introducing slackvariable s>0 to
: obtain: k=g-s=0 (this approach won't do me any numerical
: harm because of the system being low dimensioned.)
: Due to Rockafellar (augmented Lagrangian):
: f+2*theta*k+rho*k*k=f-theta*theta/rho=min
: with theta and rho being the lagrangemultiplier resp.
: the penaltyparameter.
: The iterative update is then:
: theta^(n+1)=theta^(n)+rho*g^(n)
: Is that right?
: I'd highly apppreciate any kind of help!
: TIA
: Thom@s
I'd better give you a cite:
Rockafellar, R.T.:
The multiplier method of Hestenes and Powell applied
to convex programming.
J. Opt. Theory, Appl. 12 (1973), 555-562.