Metamath Proof Explorer


Theorem nmcopexi

Description: The norm of a continuous linear Hilbert space operator exists. Theorem 3.5(i) of Beran p. 99. (Contributed by NM, 5-Feb-2006) (Proof shortened by Mario Carneiro, 17-Nov-2013) (New usage is discouraged.)

Ref Expression
Hypotheses nmcopex.1 𝑇 ∈ LinOp
nmcopex.2 𝑇 ∈ ContOp
Assertion nmcopexi ( normop𝑇 ) ∈ ℝ

Proof

Step Hyp Ref Expression
1 nmcopex.1 𝑇 ∈ LinOp
2 nmcopex.2 𝑇 ∈ ContOp
3 ax-hv0cl 0 ∈ ℋ
4 1rp 1 ∈ ℝ+
5 cnopc ( ( 𝑇 ∈ ContOp ∧ 0 ∈ ℋ ∧ 1 ∈ ℝ+ ) → ∃ 𝑦 ∈ ℝ+𝑧 ∈ ℋ ( ( norm ‘ ( 𝑧 0 ) ) < 𝑦 → ( norm ‘ ( ( 𝑇𝑧 ) − ( 𝑇 ‘ 0 ) ) ) < 1 ) )
6 2 3 4 5 mp3an 𝑦 ∈ ℝ+𝑧 ∈ ℋ ( ( norm ‘ ( 𝑧 0 ) ) < 𝑦 → ( norm ‘ ( ( 𝑇𝑧 ) − ( 𝑇 ‘ 0 ) ) ) < 1 )
7 hvsub0 ( 𝑧 ∈ ℋ → ( 𝑧 0 ) = 𝑧 )
8 7 fveq2d ( 𝑧 ∈ ℋ → ( norm ‘ ( 𝑧 0 ) ) = ( norm𝑧 ) )
9 8 breq1d ( 𝑧 ∈ ℋ → ( ( norm ‘ ( 𝑧 0 ) ) < 𝑦 ↔ ( norm𝑧 ) < 𝑦 ) )
10 1 lnop0i ( 𝑇 ‘ 0 ) = 0
11 10 oveq2i ( ( 𝑇𝑧 ) − ( 𝑇 ‘ 0 ) ) = ( ( 𝑇𝑧 ) − 0 )
12 1 lnopfi 𝑇 : ℋ ⟶ ℋ
13 12 ffvelrni ( 𝑧 ∈ ℋ → ( 𝑇𝑧 ) ∈ ℋ )
14 hvsub0 ( ( 𝑇𝑧 ) ∈ ℋ → ( ( 𝑇𝑧 ) − 0 ) = ( 𝑇𝑧 ) )
15 13 14 syl ( 𝑧 ∈ ℋ → ( ( 𝑇𝑧 ) − 0 ) = ( 𝑇𝑧 ) )
16 11 15 eqtrid ( 𝑧 ∈ ℋ → ( ( 𝑇𝑧 ) − ( 𝑇 ‘ 0 ) ) = ( 𝑇𝑧 ) )
17 16 fveq2d ( 𝑧 ∈ ℋ → ( norm ‘ ( ( 𝑇𝑧 ) − ( 𝑇 ‘ 0 ) ) ) = ( norm ‘ ( 𝑇𝑧 ) ) )
18 17 breq1d ( 𝑧 ∈ ℋ → ( ( norm ‘ ( ( 𝑇𝑧 ) − ( 𝑇 ‘ 0 ) ) ) < 1 ↔ ( norm ‘ ( 𝑇𝑧 ) ) < 1 ) )
19 9 18 imbi12d ( 𝑧 ∈ ℋ → ( ( ( norm ‘ ( 𝑧 0 ) ) < 𝑦 → ( norm ‘ ( ( 𝑇𝑧 ) − ( 𝑇 ‘ 0 ) ) ) < 1 ) ↔ ( ( norm𝑧 ) < 𝑦 → ( norm ‘ ( 𝑇𝑧 ) ) < 1 ) ) )
20 19 ralbiia ( ∀ 𝑧 ∈ ℋ ( ( norm ‘ ( 𝑧 0 ) ) < 𝑦 → ( norm ‘ ( ( 𝑇𝑧 ) − ( 𝑇 ‘ 0 ) ) ) < 1 ) ↔ ∀ 𝑧 ∈ ℋ ( ( norm𝑧 ) < 𝑦 → ( norm ‘ ( 𝑇𝑧 ) ) < 1 ) )
21 20 rexbii ( ∃ 𝑦 ∈ ℝ+𝑧 ∈ ℋ ( ( norm ‘ ( 𝑧 0 ) ) < 𝑦 → ( norm ‘ ( ( 𝑇𝑧 ) − ( 𝑇 ‘ 0 ) ) ) < 1 ) ↔ ∃ 𝑦 ∈ ℝ+𝑧 ∈ ℋ ( ( norm𝑧 ) < 𝑦 → ( norm ‘ ( 𝑇𝑧 ) ) < 1 ) )
22 6 21 mpbi 𝑦 ∈ ℝ+𝑧 ∈ ℋ ( ( norm𝑧 ) < 𝑦 → ( norm ‘ ( 𝑇𝑧 ) ) < 1 )
23 nmopval ( 𝑇 : ℋ ⟶ ℋ → ( normop𝑇 ) = sup ( { 𝑚 ∣ ∃ 𝑥 ∈ ℋ ( ( norm𝑥 ) ≤ 1 ∧ 𝑚 = ( norm ‘ ( 𝑇𝑥 ) ) ) } , ℝ* , < ) )
24 12 23 ax-mp ( normop𝑇 ) = sup ( { 𝑚 ∣ ∃ 𝑥 ∈ ℋ ( ( norm𝑥 ) ≤ 1 ∧ 𝑚 = ( norm ‘ ( 𝑇𝑥 ) ) ) } , ℝ* , < )
25 12 ffvelrni ( 𝑥 ∈ ℋ → ( 𝑇𝑥 ) ∈ ℋ )
26 normcl ( ( 𝑇𝑥 ) ∈ ℋ → ( norm ‘ ( 𝑇𝑥 ) ) ∈ ℝ )
27 25 26 syl ( 𝑥 ∈ ℋ → ( norm ‘ ( 𝑇𝑥 ) ) ∈ ℝ )
28 10 fveq2i ( norm ‘ ( 𝑇 ‘ 0 ) ) = ( norm ‘ 0 )
29 norm0 ( norm ‘ 0 ) = 0
30 28 29 eqtri ( norm ‘ ( 𝑇 ‘ 0 ) ) = 0
31 rpcn ( ( 𝑦 / 2 ) ∈ ℝ+ → ( 𝑦 / 2 ) ∈ ℂ )
32 1 lnopmuli ( ( ( 𝑦 / 2 ) ∈ ℂ ∧ 𝑥 ∈ ℋ ) → ( 𝑇 ‘ ( ( 𝑦 / 2 ) · 𝑥 ) ) = ( ( 𝑦 / 2 ) · ( 𝑇𝑥 ) ) )
33 31 32 sylan ( ( ( 𝑦 / 2 ) ∈ ℝ+𝑥 ∈ ℋ ) → ( 𝑇 ‘ ( ( 𝑦 / 2 ) · 𝑥 ) ) = ( ( 𝑦 / 2 ) · ( 𝑇𝑥 ) ) )
34 33 fveq2d ( ( ( 𝑦 / 2 ) ∈ ℝ+𝑥 ∈ ℋ ) → ( norm ‘ ( 𝑇 ‘ ( ( 𝑦 / 2 ) · 𝑥 ) ) ) = ( norm ‘ ( ( 𝑦 / 2 ) · ( 𝑇𝑥 ) ) ) )
35 norm-iii ( ( ( 𝑦 / 2 ) ∈ ℂ ∧ ( 𝑇𝑥 ) ∈ ℋ ) → ( norm ‘ ( ( 𝑦 / 2 ) · ( 𝑇𝑥 ) ) ) = ( ( abs ‘ ( 𝑦 / 2 ) ) · ( norm ‘ ( 𝑇𝑥 ) ) ) )
36 31 25 35 syl2an ( ( ( 𝑦 / 2 ) ∈ ℝ+𝑥 ∈ ℋ ) → ( norm ‘ ( ( 𝑦 / 2 ) · ( 𝑇𝑥 ) ) ) = ( ( abs ‘ ( 𝑦 / 2 ) ) · ( norm ‘ ( 𝑇𝑥 ) ) ) )
37 rpre ( ( 𝑦 / 2 ) ∈ ℝ+ → ( 𝑦 / 2 ) ∈ ℝ )
38 rpge0 ( ( 𝑦 / 2 ) ∈ ℝ+ → 0 ≤ ( 𝑦 / 2 ) )
39 37 38 absidd ( ( 𝑦 / 2 ) ∈ ℝ+ → ( abs ‘ ( 𝑦 / 2 ) ) = ( 𝑦 / 2 ) )
40 39 adantr ( ( ( 𝑦 / 2 ) ∈ ℝ+𝑥 ∈ ℋ ) → ( abs ‘ ( 𝑦 / 2 ) ) = ( 𝑦 / 2 ) )
41 40 oveq1d ( ( ( 𝑦 / 2 ) ∈ ℝ+𝑥 ∈ ℋ ) → ( ( abs ‘ ( 𝑦 / 2 ) ) · ( norm ‘ ( 𝑇𝑥 ) ) ) = ( ( 𝑦 / 2 ) · ( norm ‘ ( 𝑇𝑥 ) ) ) )
42 34 36 41 3eqtrrd ( ( ( 𝑦 / 2 ) ∈ ℝ+𝑥 ∈ ℋ ) → ( ( 𝑦 / 2 ) · ( norm ‘ ( 𝑇𝑥 ) ) ) = ( norm ‘ ( 𝑇 ‘ ( ( 𝑦 / 2 ) · 𝑥 ) ) ) )
43 22 24 27 30 42 nmcexi ( normop𝑇 ) ∈ ℝ