Operator norm

Definition (Operator norm)

For some linear map AL(Rn;Rm)A\in \mathscr{L}(\mathbb{R}^{n};\mathbb{R}^{m}), the expression ARn,Rm=sup{A(x)RmxRn=1}\lVert A \rVert _{\mathbb{R}^{n},\mathbb{R}^{m}}=\sup\{ \lVert A(\boldsymbol x) \rVert _{\mathbb{R}^{m}}\mid \lVert \boldsymbol x \rVert _{\mathbb{R}^{n}}=1 \}or for a linear functional ff f=supx:x0f(x)x\lVert f \rVert = \sup_{x:\lVert x \rVert \neq 0} \frac{|f(x)|}{\lVert x \rVert } is the operator norm of AA.

Definition (Bounded in operator norm)

Let XX be a normed vector space. We say fΓ(X;R)f\in\Gamma(X;\mathbb{R}) is bounded (in the operator norm) if M>0\exists M>0 such that f(x)MxxX|f(x)|\le M\lVert x \rVert \quad\forall x\in X

Theorem (Principle of uniform boundedness (Banach-Steinhaus))

Let (X,X),(Y,Y)(X,\lVert \cdot \rVert_{X}),(Y,\lVert \cdot \rVert_{Y}) be Banach spaces. Let (ϕα)αΛ(\phi_{\alpha})_{\alpha\in\Lambda} be a family of continuous linear mappings s.t. ϕα:XY\phi_{\alpha}:X\to Y. Assume xX:supαΛϕα(x)Y<\forall x\in X: \sup_{\alpha\in\Lambda}\lVert \phi_{\alpha}(x) \rVert_{Y}<\infty. Then supαΛϕαop<\sup_{\alpha\in\Lambda}\lVert \phi_{\alpha} \rVert _{\text{op}}<\inftywhere ϕαop=supxX1ϕα(x)Y\lVert \phi_{\alpha} \rVert _{\text{op}}=\sup_{\lVert x \rVert _{X}\le 1}\lVert \phi_{\alpha}(x) \rVert _{Y}

Linked from