What is the proximal operator for $f(\mathbf{x})=-\log(x_1x_2\dots x_n)$, $\mathbf{x}>0$ and its optimal solution?
It is easy to find the former,
$P_{f}(x)=\arg \min_{\mathbf{y}>0}-\log(y_1y_2\dots y_n)+\frac{1}{2}\|\mathbf{y}-\mathbf{x}\|^2_2,$
but solving it seems to be a different ball game. Can I get some help on this?
Your problem is separable in the coordinates:
So you can consider the one-dimensional case (i.e $n=1$), which has an algebraic solution obtainable trivially by applying the constructs detailed here to $1 \times 1$ positive-definite matrices (i.e positive scalars!).