Normed Space Def'n . Norm Let V V V be a vector space over R \mathbb{R} R , a norm ∥ ⋅ ∥ \|\cdot\| ∥ ⋅ ∥ over V V V is a function V → R + V\rightarrow \mathbb{R}^+ V → R + s.t.
positive definite: ∥ x ∥ ≥ 0 ∧ ∥ x ∥ = 0 I F F x ≡ 0 ∈ V \|x\|\geq 0 \land \|x\| = 0 IFF x\equiv 0 \in V ∥ x ∥ ≥ 0 ∧ ∥ x ∥ = 0 I FF x ≡ 0 ∈ V homogeneous ∥ a x ∥ = ∣ a ∣ ∥ x ∥ . ∀ a ∈ R \|ax\| = |a|\|x\|. \forall a\in \mathbb{R} ∥ a x ∥ = ∣ a ∣∥ x ∥.∀ a ∈ R triangular inequality ∥ x + y ∥ = ∥ x ∥ + ∥ y ∥ \|x+y\| = \|x\| + \|y\| ∥ x + y ∥ = ∥ x ∥ + ∥ y ∥ Example: Euclidean space over R n \mathbb{R}^n R n ∥ x ∥ 2 = ∑ ∣ x k ∣ 2 \|x\|_2 = \sqrt{\sum |x_k|^2} ∥ x ∥ 2 = ∑ ∣ x k ∣ 2 - positive definite ∣ x k ∣ 2 ≥ 0 ⇒ ∑ ∣ x k ∣ 2 ≥ 0 |x_k|^2 \geq 0 \Rightarrow \sqrt{\sum |x_k|^2} \geq 0 ∣ x k ∣ 2 ≥ 0 ⇒ ∑ ∣ x k ∣ 2 ≥ 0 ∥ x ∥ = 0 ⇒ x = 0 \|x\| = 0\Rightarrow x = 0 ∥ x ∥ = 0 ⇒ x = 0 - homogeneous ∥ a x ∥ = ∑ ∣ a x k ∣ 2 = ∣ a ∣ 2 ∑ ∣ x k ∣ 2 = ∣ a ∣ ∥ x ∥ \|ax\| = \sqrt{\sum |ax_k|^2} = \sqrt{|a|^2\sum |x_k|^2} = |a|\|x\| ∥ a x ∥ = ∑ ∣ a x k ∣ 2 = ∣ a ∣ 2 ∑ ∣ x k ∣ 2 = ∣ a ∣∥ x ∥ - Tri-ineq By Cauchy Swartz Inequality ∥ x + y ∥ 2 = ∑ ∣ x k + y k ∣ 2 ≤ ∑ ∣ x k ∣ 2 + ∣ y k ∣ 2 = ∑ ∣ x k ∣ 2 + ∑ ∣ y k ∣ 2 = ∥ x ∥ 2 + ∥ y ∥ 2 \|x+y\|^2 = \sum |x_k + y_k|^2 \leq \sum |x_k|^2 + |y_k|^2 = \sum |x_k|^2 + \sum|y_k|^2 = \|x\|^2 + \|y\|^2 ∥ x + y ∥ 2 = ∑ ∣ x k + y k ∣ 2 ≤ ∑ ∣ x k ∣ 2 + ∣ y k ∣ 2 = ∑ ∣ x k ∣ 2 + ∑ ∣ y k ∣ 2 = ∥ x ∥ 2 + ∥ y ∥ 2
Example: Some well-known norms Some norms are
p-norm ∥ x ∥ p : = ( ∑ ∣ x k ∣ p ) p − 1 , p ≥ 1 \|x\|_p:=(\sum |x_k|^p)^{p^{-1}}, p\geq 1 ∥ x ∥ p := ( ∑ ∣ x k ∣ p ) p − 1 , p ≥ 1 Lp-norm ∥ f ∥ L p : = ( ∫ S f ( x ) p d x ) p − 1 , p ≥ 1 \|f\|_{L_p} := (\int_S f(x)^p dx)^{p^{-1}}, p\geq 1 ∥ f ∥ L p := ( ∫ S f ( x ) p d x ) p − 1 , p ≥ 1 is a norm over C : = C:= C := the set of all continuous functions sup-norm ∥ f ∥ ∞ : = s u p { ∣ f ( x ) ∣ : x ∈ S } \|f\|_\infty:= sup\{|f(x)|:x\in S\} ∥ f ∥ ∞ := s u p { ∣ f ( x ) ∣ : x ∈ S } is a norm over C b ( S ) : = C_b(S):= C b ( S ) := the set of all continuous bounded functions or over C ( S ) C(S) C ( S ) if S S S is compact. Example: sup-norm Claim . sup-norm is a norm proof .
∣ f ( x ) ∣ ≥ 0 |f(x)| \geq 0 ∣ f ( x ) ∣ ≥ 0 , f ( x ) : = 0 ⇒ ∥ f ∥ ∞ = 0 , ∥ f ∥ ∞ = 0 ⇒ ∣ f ( x ) ∣ ≤ 0 ⇒ f ( x ) = 0 f(x):= 0 \Rightarrow \|f\|_\infty = 0, \|f\|_\infty=0\Rightarrow |f(x)|\leq 0\Rightarrow f(x)=0 f ( x ) := 0 ⇒ ∥ f ∥ ∞ = 0 , ∥ f ∥ ∞ = 0 ⇒ ∣ f ( x ) ∣ ≤ 0 ⇒ f ( x ) = 0 ∥ a f ∥ ∞ = s u p ∣ a f ( S ) ∣ = ∣ a ∣ s u p ∣ f ( S ) ∣ = ∣ a ∣ ∥ f ∥ ∞ \|af\|_\infty = sup|af(S)| = |a|sup|f(S)|= |a|\|f\|_\infty ∥ a f ∥ ∞ = s u p ∣ a f ( S ) ∣ = ∣ a ∣ s u p ∣ f ( S ) ∣ = ∣ a ∣∥ f ∥ ∞ ∥ f + g ∥ = s u p ( ∣ f + g ∣ ) ≤ s u p ( ∣ f ∣ + ∣ g ∣ ) ≤ s u p ∣ f ∣ + s u p ∣ g ∣ = ∥ f ∥ + ∥ g ∥ \|f+g\|=sup(|f+g|) \leq sup(|f| + |g|) \leq sup|f| + sup|g| = \|f\| + \|g\| ∥ f + g ∥ = s u p ( ∣ f + g ∣ ) ≤ s u p ( ∣ f ∣ + ∣ g ∣ ) ≤ s u p ∣ f ∣ + s u p ∣ g ∣ = ∥ f ∥ + ∥ g ∥ C k ( S ) : = C^k(S):= C k ( S ) := The set of all real number functions whose k-first derivative exists and continuous
Some norms are defined on C k C^k C k , such as ∥ f ∥ ′ : = max ∥ f ′ ∥ ∞ ; ∥ f ∥ C k = ∑ ∥ f ( n ) ∥ ∞ \|f\|':= \max\|f'\|_\infty; \|f\|_{C^k} = \sum \|f^{(n)}\|_\infty ∥ f ∥ ′ := max ∥ f ′ ∥ ∞ ; ∥ f ∥ C k = ∑ ∥ f ( n ) ∥ ∞
Remark C ∞ C^\infty C ∞ a.k.a. smooth are normed space that obey completeness
Topology of normed-spaces Thrm 1. The set C : = { f : [ 0 , 1 ] → R : f ( 0 ) = 1 } C:=\{f:[0,1]\rightarrow \mathbb{R}: f(0)=1\} C := { f : [ 0 , 1 ] → R : f ( 0 ) = 1 } is closed in ( C ( [ 0 , 1 ] ) , ∥ f ∥ ∞ : = s u p x ∈ [ 0 , 1 ] f ( x ) ) (C([0,1]), \|f\|_\infty:= sup_{x\in [0,1]}f(x)) ( C ([ 0 , 1 ]) , ∥ f ∥ ∞ := s u p x ∈ [ 0 , 1 ] f ( x )) .
proof . Take a sequence g n ∈ C g_n \in C g n ∈ C s.t. ∥ g n − g ∥ → 0 \|g_n - g\| \rightarrow 0 ∥ g n − g ∥ → 0 . Then, consider
∣ g ( 0 ) − 1 ∣ ≤ ∣ g ( 0 ) − g n ( 0 ) ∣ + ∣ g n ( 0 ) − 1 ∣ ≤ ∣ g ( 0 ) − g n ( 0 ) ∣ ≤ ∥ g n − g ∥ = 0 ⇒ g ( 0 ) → 1 ⇒ g ( 0 ) = 1 ⇒ g ∈ C \begin{align*} |g(0) - 1| &\leq |g(0) - g_n(0)| + |g_n(0) -1|\\ &\leq |g(0)-g_n(0)| \leq \|g_n-g\|=0\\ &\Rightarrow g(0)\rightarrow 1\Rightarrow g(0) = 1\Rightarrow g\in C \end{align*} ∣ g ( 0 ) − 1∣ ≤ ∣ g ( 0 ) − g n ( 0 ) ∣ + ∣ g n ( 0 ) − 1∣ ≤ ∣ g ( 0 ) − g n ( 0 ) ∣ ≤ ∥ g n − g ∥ = 0 ⇒ g ( 0 ) → 1 ⇒ g ( 0 ) = 1 ⇒ g ∈ C Thrm 2. Let A : = { f ∈ [ 0 , 1 ] : f ( x ) > 0 , ∥ f ′ ∥ ∞ < 2 } A:=\{f\in [0,1]: f(x) > 0, \|f'\|_\infty < 2\} A := { f ∈ [ 0 , 1 ] : f ( x ) > 0 , ∥ f ′ ∥ ∞ < 2 } is open in C 1 ( [ 0 , 1 ] ) C^1([0,1]) C 1 ([ 0 , 1 ]) wrt ∥ f ∥ ∞ , C 1 : = ∥ f ∥ ∞ + ∥ f ′ ∥ ∞ \|f\|_{\infty, C^1}:=\|f\|_\infty + \|f'\|_\infty ∥ f ∥ ∞ , C 1 := ∥ f ∥ ∞ + ∥ f ′ ∥ ∞
proof . Take any g ∈ A g\in A g ∈ A . Since g ( x ) > 0 g(x) > 0 g ( x ) > 0 , by EVT on [ 0 , 1 ] [0,1] [ 0 , 1 ] , take δ 1 \delta_1 δ 1 s.t. g ( x ) > δ 1 > 0 g(x)> \delta_1 > 0 g ( x ) > δ 1 > 0 . Since g ′ ( x ) < 2 g'(x) < 2 g ′ ( x ) < 2 , by EVT on [ 0 , 1 ] [0,1] [ 0 , 1 ] , take δ 2 \delta_2 δ 2 s.t. g ′ ( x ) < 2 − δ 2 < 2 g'(x)< 2-\delta_2 < 2 g ′ ( x ) < 2 − δ 2 < 2 . Take δ = min ( δ 1 , δ 2 ) / 2 \delta = \min(\delta_1, \delta_2)/ 2 δ = min ( δ 1 , δ 2 ) /2 .
∥ h − g ∥ ∞ < δ ⟹ ∣ h ( x ) − g ( x ) ∣ < δ ⟹ h ( x ) > g ( x ) − δ > δ 1 / 2 > 0 ∥ h ′ − g ′ ∥ ∞ < δ ⟹ ∣ h ′ ( x ) − g ′ ( x ) ∣ < δ ⟹ h ′ ( x ) < g ′ ( x ) + δ < 2 − δ 2 + δ 2 / 2 < 2 \begin{align*} \|h-g\|_\infty < \delta &\implies |h(x)-g(x)|< \delta \\ &\implies h(x) > g(x)-\delta > \delta_1/2 > 0\\ \|h'-g'\|_\infty < \delta &\implies |h'(x) - g'(x)| < \delta\\ &\implies h'(x) < g'(x)+\delta < 2-\delta_2 + \delta_2/2 < 2 \end{align*} ∥ h − g ∥ ∞ < δ ∥ h ′ − g ′ ∥ ∞ < δ ⟹ ∣ h ( x ) − g ( x ) ∣ < δ ⟹ h ( x ) > g ( x ) − δ > δ 1 /2 > 0 ⟹ ∣ h ′ ( x ) − g ′ ( x ) ∣ < δ ⟹ h ′ ( x ) < g ′ ( x ) + δ < 2 − δ 2 + δ 2 /2 < 2 Thrm 3. C c ( R ) : = C_c(\mathbb{R}):= C c ( R ) := space of compactly supported function on reals, C c ( R ) , f ∈ C c ( R ) C_c(\mathbb R), f\in C_c(\mathbb R) C c ( R ) , f ∈ C c ( R ) if ∃ M > 0 \exists M > 0 ∃ M > 0 s.t. f ( x ) = 0 , ∀ ∣ x ∣ > M f(x)=0, \forall |x| > M f ( x ) = 0 , ∀∣ x ∣ > M
Claim C c ( R ) C_c(\mathbb R) C c ( R ) is not complete wrt ∥ f ∥ ∞ \|f\|_\infty ∥ f ∥ ∞
proof . WTF a Cauchy sequence f n ∈ C c ( R ) f_n \in C_c(\mathbb R) f n ∈ C c ( R ) s.t. f m → f ∉ C c ( R ) f_m \rightarrow f\not\in C_c(\mathbb R) f m → f ∈ C c ( R ) . Take f n ( x ) = 1 − ( x / n ) 2 1 + x 2 I ( ∣ x ∣ ≤ n ) f_n(x) = \frac{1-(x/n)^2}{1+x^2}\mathbb I (|x|\leq n) f n ( x ) = 1 + x 2 1 − ( x / n ) 2 I ( ∣ x ∣ ≤ n ) and we can show that f n f_n f n is Cauchy wrt ∥ ⋅ ∥ ∞ \|\cdot\|_\infty ∥ ⋅ ∥ ∞
wlog, assume n > m n > m n > m , Consider
∥ f n − f m ∥ = sup x ∈ [ − n , n ] ∣ 1 − ( x / n ) 2 1 + x 2 − f m ( x ) ∣ \|f_n - f_m\| = \sup_{x\in [-n,n]} |\frac{1-(x/n)^2}{1+x^2} - f_m(x)| ∥ f n − f m ∥ = x ∈ [ − n , n ] sup ∣ 1 + x 2 1 − ( x / n ) 2 − f m ( x ) ∣ Suppose that x ∈ [ − m , m ] x\in [-m,m] x ∈ [ − m , m ] ,
sup x ∈ [ − n , n ] ∣ 1 − ( x / n ) 2 1 + x 2 − 1 − ( x / m ) 2 1 + x 2 ∣ = ∣ x 2 1 + x 2 ∣ ∣ n − 2 − m − 2 ∣ → ∣ 1 ∣ ∣ 0 ∣ = 0 \sup_{x\in [-n,n]} |\frac{1-(x/n)^2}{1+x^2} - \frac{1-(x/m)^2}{1+x^2}| = |\frac{x^2}{1+x^2}||n^{-2} - m^{-2}| \rightarrow |1||0|=0 x ∈ [ − n , n ] sup ∣ 1 + x 2 1 − ( x / n ) 2 − 1 + x 2 1 − ( x / m ) 2 ∣ = ∣ 1 + x 2 x 2 ∣∣ n − 2 − m − 2 ∣ → ∣1∣∣0∣ = 0 Suppose that x ∈ [ − n , − m ) ∪ ( m , n ] x\in [-n, -m)\cup (m, n] x ∈ [ − n , − m ) ∪ ( m , n ] ,
sup x ∈ [ − n , n ] ∣ 1 − ( x / n ) 2 1 + x 2 ∣ ≤ ∣ 1 1 + x 2 ∣ ≤ ∣ 1 1 + m 2 ∣ → 0 \sup_{x\in [-n,n]} |\frac{1-(x/n)^2}{1+x^2}|\leq |\frac{1}{1+x^2}|\leq |\frac{1}{1+m^2}|\rightarrow 0 x ∈ [ − n , n ] sup ∣ 1 + x 2 1 − ( x / n ) 2 ∣ ≤ ∣ 1 + x 2 1 ∣ ≤ ∣ 1 + m 2 1 ∣ → 0 Therefore, take N = ϵ 1 / 2 N = \epsilon^{1/2} N = ϵ 1/2 we can prove Cauchy.
However, f n ( x ) = 1 − ( x / n ) 2 1 + x 2 → 1 1 + x 2 → 0 f_n(x)=\frac{1-(x/n)^2}{1+x^2}\rightarrow \frac{1}{1+x^2}\rightarrow 0 f n ( x ) = 1 + x 2 1 − ( x / n ) 2 → 1 + x 2 1 → 0 is not compactly supported
Therefore, only compact in metric space → \rightarrow → closed and bounded but not the converse.
January 11, 2025 January 9, 2023