Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

i have a non linear optimization problem and i am trying to solve it via lagrange multipliers.

I have 2 constraints, one with a lower bound that requires $f>0$ and one with an upper bound that requires $f<1$. The problem is very complicated so i solve it in 2 steps, using 2 different lagrangians, one for each constraint. So i search for a combination from the results that satisfies my constraints.

In some cases, because i have vectors as inputs, i did not manage to to get a solution $(x,y)$ that satisfies both $f>0$ and $f<1$. Is it because i split the problem? Thanks in advance

P.S i tried not to split the problem, and have one lagrangian with 2 constraints but because of the complexity of maths, maple and matlab did not return values. Thanks in advance.

share|cite|improve this question

Interesting, if you have an alaytic $f$, it would help if you could paste it here. I don't think it is any more complex to handle another boundry constraint...

Maybe the problem is an anemic solver... can you try representing the problem differently? Let $f$ be your optimization function, then you can explicitly constrain $0 < f < 1$ (instead of using a complex expression which a dumb solver may not identify with the optimization function) and then minimize $f$.

I.e. if you are minimizing some complex expression, constrain $f=\{\text{that expression}\}$ and add a constraint $0 < f < 1$ and then minimize $f$...

share|cite|improve this answer
Thanks for your answer.My optimization function is ,in fact a least square problem,so i want to minimize the function (estim-data)^2,so that f>0 and f<1.The function f,is parameterised by x and y,and so is estim. – user86377 Jul 15 '13 at 17:55
@user86377 I don't understand - are you minimizing $$ \sum_{k=1}^N \left( f \left(x,y,\vec{v}_k \right) - y_k \right)^2 $$ for some fixed $\vec{v}_k, y_k, f$ over parameters $x,y$ and want to constrain that $f\left(x,y,\vec{v}_k \right) \in (0,1) \forall x,y,\vec{v}_k$? – gt6989b Jul 17 '13 at 15:47
I want to fit a curve.Let d, be the points of that curve.So I want to minimize the nonlinear error $$ \sum_{k=1}^N \left( f \left(x,y \right) - d_k\right)^2 $$ s.t $$ g\left(x,y \right) \in (0,1) $$ with the optimal combination of (x,y) – user86377 Jul 18 '13 at 8:26
@user86377 You mean $f(x,y) \in (0,1)$ or $g$ is realy indended to be a different function? – gt6989b Jul 18 '13 at 13:04
No, g is a different function that i want to bound between 0 and 1, while minimizing the error with f. g and f are functions with some exponentials, but they are different. – user86377 Jul 18 '13 at 13:37

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.