In the setting of Routine Facts Used Without Citation: Arithmetic, Maps, Predicativity and Indices , let n ∈ N n\in\mathbb{N} n ∈ N , with the square real matrices M n ( R ) \mathcal{M}_{n}(\mathbb{R}) M n ( R ) , invertible matrices C C C and their inverses C − 1 C^{-1} C − 1 , the operator norm ∥ C ∥ \lVert C\rVert ∥ C ∥ of a matrix C C C , and the distance d n d_{n} d n , a metric on M n ( R ) \mathcal{M}_{n}(\mathbb{R}) M n ( R ) by The Real Matrices with the Operator-Norm Distance Form a Complete Metric Space with Entrywise Convergence §metric . Let G L n ( R ) = { C ∈ M n ( R ) : C is invertible } \mathrm{GL}_{n}(\mathbb{R})=\{C\in\mathcal{M}_{n}(\mathbb{R}):C\text{ is invertible}\} GL n ( R ) = { C ∈ M n ( R ) : C is invertible } be the set of the invertible matrices in M n ( R ) \mathcal{M}_{n}(\mathbb{R}) M n ( R ) of The Invertible Matrices and the Positive Semidefinite Matrices Form Sets §invertible .
¶ Let A ∈ M n ( R ) A\in\mathcal{M}_{n}(\mathbb{R}) A ∈ M n ( R ) be invertible. If B ∈ M n ( R ) B\in\mathcal{M}_{n}(\mathbb{R}) B ∈ M n ( R ) and ∥ A − 1 ∥ d n ( A , B ) < 1 \lVert A^{-1}\rVert\,d_{n}(A,B)<1 ∥ A − 1 ∥ d n ( A , B ) < 1 , then B B B is invertible and
d n ( A − 1 , B − 1 ) ≤ ∥ A − 1 ∥ 2 d n ( A , B ) 1 − ∥ A − 1 ∥ d n ( A , B ) . d_{n}(A^{-1},B^{-1})\le\frac{\lVert A^{-1}\rVert^{2}\,d_{n}(A,B)}{1-\lVert A^{-1}\rVert\,d_{n}(A,B)} . d n ( A − 1 , B − 1 ) ≤ 1 − ∥ A − 1 ∥ d n ( A , B ) ∥ A − 1 ∥ 2 d n ( A , B ) .
¶ G L n ( R ) \mathrm{GL}_{n}(\mathbb{R}) GL n ( R ) is open in ( M n ( R ) , d n ) (\mathcal{M}_{n}(\mathbb{R}),d_{n}) ( M n ( R ) , d n ) .
¶ Let A ∈ M n ( R ) A\in\mathcal{M}_{n}(\mathbb{R}) A ∈ M n ( R ) be invertible. For every real ε > 0 \varepsilon>0 ε > 0 there is a real δ > 0 \delta>0 δ > 0 such that every B ∈ M n ( R ) B\in\mathcal{M}_{n}(\mathbb{R}) B ∈ M n ( R ) with d n ( A , B ) < δ d_{n}(A,B)<\delta d n ( A , B ) < δ is invertible and satisfies d n ( A − 1 , B − 1 ) < ε d_{n}(A^{-1},B^{-1})<\varepsilon d n ( A − 1 , B − 1 ) < ε .
¶ Let d G L d_{\mathrm{GL}} d GL be the restriction of d n d_{n} d n to G L n ( R ) × G L n ( R ) \mathrm{GL}_{n}(\mathbb{R})\times\mathrm{GL}_{n}(\mathbb{R}) GL n ( R ) × GL n ( R ) , a metric on G L n ( R ) \mathrm{GL}_{n}(\mathbb{R}) GL n ( R ) by Subsets of a Metric Space: the Restricted Metric, Sequences, Limits in Closed Subsets and Completeness of Closed Subsets §restriction . There is exactly one map f : G L n ( R ) → M n ( R ) f:\mathrm{GL}_{n}(\mathbb{R})\to\mathcal{M}_{n}(\mathbb{R}) f : GL n ( R ) → M n ( R ) with f ( B ) = B − 1 f(B)=B^{-1} f ( B ) = B − 1 for every B ∈ G L n ( R ) B\in\mathrm{GL}_{n}(\mathbb{R}) B ∈ GL n ( R ) , the map B ↦ B − 1 B\mapsto B^{-1} B ↦ B − 1 ; and f f f , from ( G L n ( R ) , d G L ) (\mathrm{GL}_{n}(\mathbb{R}),d_{\mathrm{GL}}) ( GL n ( R ) , d GL ) to ( M n ( R ) , d n ) (\mathcal{M}_{n}(\mathbb{R}),d_{n}) ( M n ( R ) , d n ) , is continuous .