0
$\begingroup$

Under what conditions will a solution $y \in \Re^n$ exist for this system of nonlinear equations? If it exists, will it be unique?

$$ \mu_k = \int_0^\infty g_k(x) \, f(x, y) \, dx \qquad \forall \, k = 1, \dotsm, n $$

where $\mu_k$ are constants and all functions here are continuous.

If we cannot conclude anything in general, then consider a specific case in probability theory:

$$ \begin{align} \mu_k &= \frac{1}{Z(y)} \int_0^\infty g_k(x) \, e^{f(x, y)} \, dx \qquad \forall \, k = 1, \dotsm, n \\[1ex] Z(y) &= \int_0^\infty e^{f(x, y)} \, dx \end{align} $$

where $Z(y)$ is a normalizing constant while $g_k$ and $f$ are piecewise linear in $x$. Their exact forms depend on some data, but the following is true regardless of the data:

$$ \lim_{x \rightarrow \infty} f(x, y) = \begin{cases} \infty & \text{if } \; c^T y > 0 \\ -\infty & \text{if } \; c^T y < 0 \end{cases} $$

where $c \in \Re^n$ is constant.

$\endgroup$
5
  • $\begingroup$ Do you have a model case? There could be more different approaches, according to the different situations. $\endgroup$ Commented Dec 24, 2016 at 10:22
  • $\begingroup$ For what system of nonlinear equations? $\endgroup$ Commented Dec 24, 2016 at 17:35
  • 1
    $\begingroup$ @PietroMajer I've added a model case to the question. $\endgroup$ Commented Dec 24, 2016 at 17:36
  • $\begingroup$ @AmirSagiv Why downvote? The system of nonlinear equations is already stated in the question. By the way, I'm not trying to solve a differential equation; I'm solving for $y \in \Re^n$ $\endgroup$ Commented Dec 24, 2016 at 17:44
  • $\begingroup$ got it, canceled the downvote $\endgroup$ Commented Dec 24, 2016 at 17:54

0

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.