technical note

§ Sequence That Converges Weakly but Not Strongly In lpl^plp.

created 2023-08-06 · last edited 2023-09-22
  • Consider the sequence e1=(1,0,… )e_1 = (1, 0, \dots)e1​=(1,0,…), e2≡(0,1,… )e_2 \equiv (0, 1, \dots)e2​≡(0,1,…), and in general ei[j]=δije_i[j] = \delta_i^jei​[j]=δij​.
  • Recall that to check weak convergence, it suffices to check on a basis of the dual space.
  • We check on the basis πj(x)↦x[j]\pi_j (x) \mapsto x[j]πj​(x)↦x[j].
  • Clearly, on such a basis, we see that lim⁡n→∞en[j]→0\lim_{n \to \infty} e_n[j] \to 0limn→∞​en​[j]→0, because after n>jn > jn>j, the sequence will be forever zero.
  • However, see that this sequence does not strongly converge, since the basis vectors eie_iei​ cannot be cauchy, since ∣∣ei−ej∣∣=(2)||e_i - e_j|| = \sqrt(2)∣∣ei​−ej​∣∣=(​2) when i≠ji \neq ji=j.
  • The intuition is that weak convergence can only see converge "in a finite subspace", since we are considering what happens with bounded linear functionals.
  • Thus, a sequence can appear to converge when restricting attention to any finite region of space, but cannot strongly converge.
❦
Newer ৪ Blog ৪ Older