I would like to ask the following question.
I have an aggregated function, $$ \mu(x_1,x_2) = \|\nabla f_2\| f_1(\vec{x}) + \|\nabla f_1\| f_2(\vec{x}), $$
where the norm of the gradients are also functions of $x_1$ and $x_2$.
I would like to get the equation describing the curve for which $\mu = 0$.
As an example, for the functions,
$$ f_1 = (x_1 - 3)^2 + (x_2 -3)^2 $$ and $$ f_2 = \frac{1}{2}(x_1 - 1)^2 + (x_2 - 1)^2 $$
I get,
$$ \mu(x_1,x_2) = \left[ (x_1 - 1)^2 + 4(x_2 - 1)^2 \right]^{\frac{1}{2}} \left[ (x_1 - 3)^2 + (x_2 -3)^2 \right] + 2\left[ ((x_1 - 3)^2 + (x_2 - 3)^2\right]^{\frac{1}{2}} \left[\frac{1}{2}(x_1 - 1)^2 + (x_2 - 1)^2\right] = 0, $$
and I need some function describing the solution, like $\gamma(t)$ or $\gamma(x_1,x_2)$. And it should be goning from $(1,1)$ to $(3,3)$. It would be nice to get a general approach explained, as I want to do this for other more complex functions and with more dimensions. However, I would be happy with the solution to this example since I use it in the report.
(I'am not sure if it is possible or easy, I'am just at a lost and any comment helps.)

I want to find that curve from $p_1$ to $p_2$ at the lowest point, which I suppost, would be the solution to,
$$ \mu = \frac{1}{|\nabla f_1(\vec{x})|} \nabla f_1(\vec{x}) + \frac{1}{|\nabla f_2(\vec{x})|} \nabla f_2(\vec{x}) = 0 $$
or is it not possible to use the norms depending on $\vec{x}$?
– DI Kappelle Nov 28 '18 at 16:58