In a paper I'm reading, they minimize a function $g(\alpha)$ subject to something like:
$a(\alpha)=b(\alpha)=c$ $\quad$ (Eq. 6 in the paper)
and
$d(\alpha)=e(\alpha)=f$ $\quad$ (Eq. 7 in the paper)
They skip the details of the minimization, but they do say they used Lagrange multipliers and give a closed form expression for the minimizer as $\alpha(\lambda_1, \lambda_2)$.
My question is, how can I reduce the constraints above to only two equations, so the minimizer will end up being a function of only two lambdas?
I thought about:
$a(\alpha) - 2b(\alpha) +c = 0$
and
$d(\alpha) - 2e(\alpha) +f = 0$
but I realize these don't necessarily imply the original constraints.
One way: \begin{align*} (a(\alpha) - c)^2 + (b(\alpha) - c)^2 = 0 &\text{ and} \\ (d(\alpha) - f)^2 + (e(\alpha) - f)^2 = 0 &\text{.} \end{align*}
This works if everything in sight is a real number or a real valued function. It does not work in the complex setting.