zoukankan      html  css  js  c++  java
  • Comparison of several types of convergence

    In functional analysis, several types of convergence are defined, namely,

    • strong convergence for elements in normed linear space.
    • weak convergence for elements in normed linear space, which is defined via the assistance of the dual space.
    • weak-* convergence for linear functionals in the strong dual space of a normed linear space.
    • pointwise convergence for linear operators.

    This post summarizes their definitions and shows the differences.

    1. Definition (Strong convergence) Let (X) be a normed linear space and ((x_l)_{l in mathbb{N}}) be a sequence in (X). Then ((x_l)_{l in mathbb{N}}) converges (strongly) to (x in X) if

      [
      lim_{l ightarrow infty} orm{x_l - x}_X = 0.
      ]

      It can be seen that the strong convergence is just the convergence with respect to the “distance between points”, or more generally, the so-called “norm” defined for a linear space, which is what we have been familiar with in fundamental calculus.

    2. Definition (Weak convergence) Let (X) be a Banach space and (X’) be its dual space. The sequence ((x_l)_{l in mathbb{N}}) in (X) converges weakly to (x in X) if

      [
      lim_{l ightarrow infty} abs{f(x_l) - f(x)} = 0 quad (forall f in X’).
      ]

      We can see that the convergence here is called weak, because it is not directly based on point distance in the original space (X), but the evaluation of an arbitrary functional in the dual space on the sequence.

      It is easy and natural to see that the strong convergence implies weak convergence because of the continuity of the linear functional (f in X’):

      [
      abs{f(x_l) - f(x)} = abs{f(x_l - x)} leq orm{f}_{X’} orm{x_l - x}_X.
      ]

    3. Definition (Pointwise convergence) Let (X) and (Y) be normed spaces. The sequence of bounded linear operators ((T_l)_{l in mathbb{N}} subset L(X, Y)) converges to (T in L(X, Y)) if

      [
      lim_{l ightarrow infty} orm{T_l x - T x}_Y = 0 quad (forall x in X).
      ]

      The pointwise convergence is used to describe the convergence of operators at each point in (X). A more strict convergence for operators is uniform convergence, which means the convergence speeds of ((T_l x)_{l in mathbb{N}}) at different points (x) in (X) are comparable. It is also easy to see that the strong convergence of ((T_l)_{l in mathbb{N}}) implies pointwise convergence.

    4. Definition (Weak-* convergence) Let (X_s’) be the strong dual space of the normed linear space (X). The linear functional sequence ((T_l)_{l in mathbb{N}}) converges to (T) in (X_s’) if

      [
      lim_{l ightarrow infty} abs{T_l x - T x} = 0 quad (forall x in X).
      ]

      The weak-* convergence can be considered as a special case of pointwise convergence with the difference that the linear operators become linear functionals and the dual space (X’) of (X) is assigned with the strong topology.

  • 相关阅读:
    iOS之使用AFN进行序列化处理(5)
    iOS之AFN文件上传(4)
    iOS之AFN使用技巧(3)
    iOS之AFN框架基本使用(2)
    iOS之AFN简介(1)
    iOS之JSON解析和视频的简单播放
    iOS之带有边框的圆形图片裁剪
    SQL limit offset
    Ubuntu安装pycharm, No JDK found. Please validate either PYCHARM_JDK, JDK_HOME or JAVA_HOME environment variable points to valid JDK installation
    scrapy项目运行报错: exceptions.ImportError: No module named win32api
  • 原文地址:https://www.cnblogs.com/peabody/p/7696131.html
Copyright © 2011-2022 走看看