Something similar to the Minkowski's lattice theorem for the locally compact groups (problem with an exercise)

20 Views Asked by At

There is the following problem in the book "Fourier Analysis on Number Fields" by D. Ramakrishnan and R. Valenza:
Let $G$ be a locally compact abelian subgroup with Haar measure $\mu$, and $\Gamma$ a subgroup of $G$ (not necessarily closed) such that $G=\Gamma+\Phi$ for some compact subset $\Phi\subset G$. If $X$ is a compact subset of $G$ such that $\mu(X)>\mu(\Phi)$ then there exist two different elements $x_1,x_2\in X$ such that $x_1-x_2\in \Gamma$.
I was only able to show this for $\Gamma$ discrete (or countable). I put a sketch of my partial solution below (I wanted to make it hidden, to not suggest an approach that may not possibly work in general, but I am doing something wrong with the command). Any hints or solutions are welcome.

Assuming the contrary, all the translations $\gamma+X$, for $\gamma\in\Gamma$, are pairwise disjoint. Let $$S=\{\gamma\in\Gamma:(\gamma+X)\cap \Phi\neq \emptyset\}=\Gamma\cap (X-\Phi).$$ The set $S$ is finite since $\Gamma$ is discrete and $X-\Phi$ is compact. Setting $$Y_\gamma=(\gamma+X)\cap\Phi,\ \ \ Z_\gamma=(Y_\gamma-\gamma)\subset X,$$ we obtain $$\bigsqcup_S Y_\gamma \subset \Phi\implies \sum_S\mu(Y_\gamma)\leq\mu(\Phi)$$ $$\bigcup_S Z_\gamma=X\implies \sum_S\mu(Z_\gamma)\geq \mu(X)$$ which gives the contradiction since $\mu(Y_\gamma)=\mu(Z_\gamma)$.