are upper right triangular matrices a locally compact group?

103 Views Asked by At

Let

$$G=\left\{\begin{pmatrix}a & b\\0& 1\end{pmatrix}:a>0,b\in \mathbb{R}\right\}$$

be endowed with matrix multiplication and the inherited topology as a subspace of $\mathbb{R}^4$. Show that $G$ is a locally compact group.

A locally compact group is a topological group whose topology is locally compact and Hausdorff. So, I'd have to show that. However, I would also have to show that $G$ is a topological group. $G$ is a group, which is easy to show; but I would need to show that $(x,y)\rightarrow xy$ and $x\rightarrow x^{-1}$ are continuous (in this case, it should be, I believe, $(w,x,y,z)$ since we are in a subspace of $\mathbb{R}^4$). Long story short, I am confused which direction to go. I'd suspect that there should be a relatively straightforward easy way to show that $G$ is a locally compact group, because from what I've seen, that is typically just assumed. I found this, but unfortunately didn't gain much insight from it.

I'd appreciate any guidance/hint. Thanks.

1

There are 1 best solutions below

6
On BEST ANSWER

You're right that showing that this group is locally compact isn't too bad. I'll assume that you're OK with the continuity of the group operations, these are just rational/linear functions so it should be clear that they are continuous.

To show that the topology is Hausdorff is inherited from the topology of $\mathbb{R}^4$. If $A,B \in \mathbb{R}^4$ then there exist open sets $A \in U,B \in V$ with $U \cap V = \emptyset$. Now instead of arbitrary elements of $\mathbb{R}^4$ assume that $A,B \in G$. The sets $U\cap G,V \cap G$ are open sets in the topology of $G$ that appropriately separate $A$ and $B$ so $G$ is Hausdorff.

Local compactness is also inherited but here is a long-winded argument. First consider the subset of $\mathbb{R}^4$, $S = \{(x,y,1,0):x,y, \in \mathbb{R}^2\}$, which is closed. This subset of $\mathbb{R}^4$ will be locally compact in its subspace topology in light of the fact that intersections of compact sets with closed sets are compact (see **). Now look at the subspace of $S$, $G$. Notice that $G$ is an open set in $S$. Let $C$ be a compact neighbourhood in $S$ of a point $A \in G$. Thus we can find an open ball $B_{\delta}(A) \cap S$ in $S$ about $A$ that is contained in $G$ and whose closure is contained in $C$. Then $cl(B_{\delta(A)} \cap S) \subset G$ is a compact set in $S$ (being closed and contained in the compact set $C$ -- note we need that $S$ is Hausdorff for this). In addition, $cl(B_{\delta(A)} \cap S)$ contains the set $B_{\delta(A)} \cap S$ which is open in $G$. To see that $cl(B_{\delta(A)} \cap S)$ is actually compact in the topology of $G$ you can revert to the open cover definition of compactness and cover it with open sets of $G$ and see if you can find a finite subcover. The compactness of $cl(B_{\delta(A)} \cap S)$ in $S$ will help you here (see ** again).

(**) Basically all you need to show is that if a set, $C$ is compact in a topological space ,$X$, and $C \subset Y \subset X$ then $C$ is compact in the subspace topology of $Y$.

Rather than go through the above mess of an argument it is easier to show local compactness constructively: for $(a,b,1,0) \in G$ the set $C = \{(x,y,1,0): x \in [\frac{a}{2},\frac{3a}{2}], y \in [y-1,y+1]\}$ is a compact neighbourhood of $G$.

In general showing something like local compactness for a matrix group shouldn't give you too much grief. Usually the group is either an open of closed set of $\mathbb{R}^n$. In this case this matrix group should be viewed as an open set of $\mathbb{R}^2$.