Is the inverse of a bounded operator always bounded , if yes how to prove it ?
-
1In which kind of spaces are you working? If you are working in Banach spaces, I think this can be a useful topic – PhDing Jan 03 '16 at 11:12
-
Related : https://math.stackexchange.com/questions/1580369 – Watson Apr 06 '16 at 10:08
3 Answers
If You consider an invertible, i.e. bijective and bounded linear operator $A:X\rightarrow Y$, between two Banach-spaces (it´s important they are complete), then as a consequence of Baires category theorem A is open (open mapping theorem) and so $A^{-1}:Y \rightarrow X$ is continuous, i.e. bounded.
- 4,353
Let T be a bounded linear operator from a normed space X on to a normed space Y. If there exists a positive b such that $||Tx||\geq b ||x|| \ \forall x\in X,$ then $T^{-1}$ exists and is bounded.
- 414
-
I See that if $T^{-1}$ exists it would be bounded, but could you give an argument for the existence of $T^{-1}$? – Aug 17 '21 at 15:49
-
In fact consider the right shift $T:\mathcal{l}^2\to \mathcal{l}^2,(x_1,x_2,x_3,\dots)\mapsto (0,x_1,x_2,\dots)$. Then $\Vert Tx\Vert = \Vert x \Vert$, but $T$ is not surjective. – Aug 18 '21 at 07:41
-
@user592521 https://www.math.cuhk.edu.hk/course_builder/1516/math4010/solution4.pdf there you go! – ARROW Feb 18 '24 at 11:54
It depends on how you define "inverse" of an operator. In general, if you don't require it be bounded, then no. For example, consider the bijection from $\mathbb{R}$ onto $(-1,1)$. But sometimes we say an operator is invertable only if we can find a bounded operator such that the composition is identity.
- 13,204
-
-
, thanks , i think i found it here http://planetmath.org/boundedinversetheorem – lucaluca Jan 03 '16 at 11:23
-
-
2But operator means linear operator and there is no bounded linear operator from $\mathbb{R}$ to $(-1,1)$. So I don't understand this answer. – Caleb Stanford Jul 30 '16 at 09:27