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

Problem statement:

I am given 3 sets of equations that govern the force $P$, and also the neutral axis, defined by two variables, the radius from the center $r$ and also the rotation degree in $\theta$.

The three simultaneous nonlinear equations I have to solve are listed below:

$$f_1(P,r,\theta)=P-p_c(r,\theta)$$ $$f_2(P,r,\theta)=e_yP-m_{cx}(r,\theta)$$ $$f_3(P,r,\theta)=e_xP-m_{cy}(r,\theta)$$

In order to obtain $P$,$r$ and $\theta$, I would have to solve for $f_1=0$, $f_2=0$ and $f_3=0$.

$e_y$ and $e_x$ are constants.

The constraints that govern the neutral axis variable are

$$-\infty<r<\infty$$ $$0\leq\theta\leq\pi$$

Additional information:

The $p_c$, $m_{cx}$ and $m_{cy}$ are at least differentiable, in the sense that we assume $\frac{\partial p_c(r,\theta)}{\partial r}$, $\frac{\partial p_c(r,\theta)}{\partial \theta}$ and all the terms inside the Jacobian matrix exist, but can only be evaluated through numerical means. As for the terms in hessian matrix, although it is in principle obtainable via numerical means, but I afraid that the error would be too big to be useful.


how to solve for these system of equations?


I am thinking about using quasi-newton method to solve this problem, but I don't know how to incorporate the constraints into the quasi-newton method formulation.

Another way is to formulate this into an optimization problem, i.e, minimize for the following $f$ function

$$f=f_1^2(P,r, \theta)+f_2^2(P,r, \theta)+f_3^2(P,r, \theta)$$

and the following constraints:


Is this a convex optimization problem?

But I don't know what are the techniques I can use in this highly nonlinear optimization problem.

share|cite|improve this question
If I'm reading correctly, you're asking us how to solve an equation where technically we still don't really even know what the equations explicitly are: there's no mention as to what $p_c, m_{cx},m_{cy}$ are defined by. You might as well ask us how to generally attack the problem $f(x)=0$ without knowing $f$. Is my understanding accurate? – anon Sep 18 '11 at 8:11
@anon, yes. We don't know what these equations explicitly are. All we know is that these terms and their first derivatives are evaluatable by numerical means. – Graviton Sep 18 '11 at 8:16
up vote 2 down vote accepted

In lack of other information about the system behavior, I'd go for minpack. It has hybrid Powell's method for non-linear equations and Levenberg-Marquardt algorithm for minimization problems, both with explicit Jacobian or its numerical evaluation. There is comprehensive documentation referenced at the wikipedia page (original docs to minpack).

There is a c++ port in eigen's non-linear optimization module. The same algorithms are used by Matlab's fsolve and scipy.optimization.fsolve.

share|cite|improve this answer

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.