Minimizing a non-linear function

43 Views Asked by At

I am trying to minimize the following equation

$$ C(\rho) = \| I-\sqrt\rho \nabla. \frac{1}{\rho} \nabla \sqrt\rho\|$$

where $I(x,y)$ and $\rho(x,y)$ are functions of x and y.

I found solution for this equation in a paper where new variables were introduced

$$ \alpha^4 = \rho$$ $$ \vec{F} = \frac{1}{\beta}\nabla \beta$$

then the equation is transformed to $ C(\rho, \alpha, \vec{F})= min$

and the solution is

$$ C = \|I+|\vec{F}|^2 - \nabla . \vec{F}\|^2 + \|\vec{F}- \frac{\nabla \alpha^2}{\alpha^2}\|^2 + \epsilon \|\vec{F}\|^2$$

where $\epsilon$ is positive real penalization parameters. The details of this derivation is not included in the paper.

I want to understand how this solution is derived.

it seems to me if I solve

$$ I-\sqrt\rho \nabla. \frac{1}{\rho} \nabla \sqrt\rho = 0 $$

It is easier because this is Poisson equation $$\sqrt\rho \nabla. \frac{1}{\rho} \nabla \sqrt\rho = I$$

$$\sqrt\rho \nabla^2 \frac{\sqrt\rho}{\rho} = I$$

Are the two solutions are similar or Am I am missing something