# Implementation of limits on the optimization results

3 views (last 30 days)
Itachi on 11 Sep 2012
Hi guys, I've want to optimize my problem using "fmincon". I want to put limits on the answers, let's say I want x1 > x2 and x2 > x3 in here:
x = fmincon(fun,x0,A,b,[],[],nonlincon)
how should I implement this kinda constraints?
I've set "A" to [-1 1 0;0 -1 1] and "b" to zeros(size(A,1),1) but I get infeasible results! Shouldn't it work?! Please help me out.
Best regards,

Oleg Komarov on 11 Sep 2012
Edited: Oleg Komarov on 11 Sep 2012
Should it be
b = zeros(size(A,1),1);
An example problem (without nonlinear constraints):
f = @(x) x(1)^2 + x(2)^2 + x(3)^2;
A = [-1 1 0
0 -1 1];
b = zeros(size(A,1),1);
X = fmincon(f,3:-1:1,A,b);

Itachi on 11 Sep 2012
Well, I'm saying in any problem if we want to reach this goal we should apply similar methodes like above, right?
Oleg Komarov on 11 Sep 2012
I showed you some syntax, not methods.
If you're alluding to the choice of X0 (the starting values for the x) then I just picked up something that did not satisfy the constraints to check afterwards that they were satisfied.
Also, you were talking about x1, x2, x3 and f is an example of a function in 3 variables. Basically, that's how you should write it.
Itachi on 11 Sep 2012
Thanks.