<?xml version="1.0"?>
<feed xmlns="http://www.w3.org/2005/Atom" xml:lang="en">
	<id>https://calculus.subwiki.org/w/index.php?action=history&amp;feed=atom&amp;title=Rate_of_convergence</id>
	<title>Rate of convergence - Revision history</title>
	<link rel="self" type="application/atom+xml" href="https://calculus.subwiki.org/w/index.php?action=history&amp;feed=atom&amp;title=Rate_of_convergence"/>
	<link rel="alternate" type="text/html" href="https://calculus.subwiki.org/w/index.php?title=Rate_of_convergence&amp;action=history"/>
	<updated>2026-04-06T08:21:14Z</updated>
	<subtitle>Revision history for this page on the wiki</subtitle>
	<generator>MediaWiki 1.41.2</generator>
	<entry>
		<id>https://calculus.subwiki.org/w/index.php?title=Rate_of_convergence&amp;diff=2476&amp;oldid=prev</id>
		<title>Vipul: Created page with &quot;==Definition==  The &#039;&#039;&#039;rate of convergence&#039;&#039;&#039; for a convergent sequence describes how quickly the terms of the sequence converge to the limit. We discuss some cases below....&quot;</title>
		<link rel="alternate" type="text/html" href="https://calculus.subwiki.org/w/index.php?title=Rate_of_convergence&amp;diff=2476&amp;oldid=prev"/>
		<updated>2014-04-26T16:18:14Z</updated>

		<summary type="html">&lt;p&gt;Created page with &amp;quot;==Definition==  The &amp;#039;&amp;#039;&amp;#039;rate of convergence&amp;#039;&amp;#039;&amp;#039; for a &lt;a href=&quot;/w/index.php?title=Convergent_sequence&amp;amp;action=edit&amp;amp;redlink=1&quot; class=&quot;new&quot; title=&quot;Convergent sequence (page does not exist)&quot;&gt;convergent sequence&lt;/a&gt; describes how quickly the terms of the sequence converge to the limit. We discuss some cases below....&amp;quot;&lt;/p&gt;
&lt;p&gt;&lt;b&gt;New page&lt;/b&gt;&lt;/p&gt;&lt;div&gt;==Definition==&lt;br /&gt;
&lt;br /&gt;
The &amp;#039;&amp;#039;&amp;#039;rate of convergence&amp;#039;&amp;#039;&amp;#039; for a [[convergent sequence]] describes how quickly the terms of the sequence converge to the limit. We discuss some cases below.&lt;br /&gt;
&lt;br /&gt;
For the definitions below, the sequence is &amp;lt;math&amp;gt;(x_n)_{n \in \mathbb{N}}&amp;lt;/math&amp;gt; and the limit is &amp;lt;math&amp;gt;L&amp;lt;/math&amp;gt;. In symbols, we have:&lt;br /&gt;
&lt;br /&gt;
&amp;lt;math&amp;gt;\lim_{n \to \infty} x_n = L&amp;lt;/math&amp;gt;&lt;br /&gt;
&lt;br /&gt;
===Steadily linear, superlinear, and sublinear convergence===&lt;br /&gt;
&lt;br /&gt;
Suppose the following holds:&lt;br /&gt;
&lt;br /&gt;
&amp;lt;math&amp;gt;\lim_{n \to \infty} \frac{|x_{n+1} - L|}{|x_n - L|} = \mu&amp;lt;/math&amp;gt;&lt;br /&gt;
&lt;br /&gt;
Clearly, &amp;lt;math&amp;gt;\mu \in [0,1]&amp;lt;/math&amp;gt; (if &amp;lt;math&amp;gt;\mu&amp;lt;/math&amp;gt; were greater than 1, the sequence couldn&amp;#039;t converge). We have three cases:&lt;br /&gt;
&lt;br /&gt;
* &amp;lt;math&amp;gt;\mu \in (0,1)&amp;lt;/math&amp;gt;, i.e., &amp;lt;math&amp;gt;0 &amp;lt; \mu &amp;lt; 1&amp;lt;/math&amp;gt;: In this case, the sequence &amp;lt;math&amp;gt;(x_n)&amp;lt;/math&amp;gt; converges to &amp;lt;math&amp;gt;L&amp;lt;/math&amp;gt; linearly, and the constant &amp;lt;math&amp;gt;\mu&amp;lt;/math&amp;gt; is called the rate of (linear) convergence. The use of the term &amp;quot;linear&amp;quot; signifies that the next error term &amp;lt;math&amp;gt;|x_{n+1} - L|&amp;lt;/math&amp;gt; is approximately a &amp;#039;&amp;#039;linear&amp;#039;&amp;#039; function of the preceding error term &amp;lt;math&amp;gt;|x_n - L|&amp;lt;/math&amp;gt;.&lt;br /&gt;
* &amp;lt;math&amp;gt;\mu = 0&amp;lt;/math&amp;gt;: In this case, the sequence &amp;lt;math&amp;gt;(x_n)&amp;lt;/math&amp;gt; converges to &amp;lt;math&amp;gt;L&amp;lt;/math&amp;gt; &amp;#039;&amp;#039;super&amp;#039;&amp;#039;linearly. In other words, a function that would approximate the behavior of the next error term in terms of the previous error term would decay faster than a linear function.&lt;br /&gt;
* &amp;lt;math&amp;gt;\mu = 1&amp;lt;/math&amp;gt;: In this case, the sequence &amp;lt;math&amp;gt;(x_n)&amp;lt;/math&amp;gt; converges to &amp;lt;math&amp;gt;L&amp;lt;/math&amp;gt; &amp;#039;&amp;#039;sub&amp;#039;&amp;#039;linearly. In other words, a function that would describe the next error term in terms of the previous one would decay slower than a linear function.&lt;br /&gt;
&lt;br /&gt;
===Convergence of a power function form===&lt;br /&gt;
&lt;br /&gt;
Suppose &amp;lt;math&amp;gt;q &amp;gt; 1&amp;lt;/math&amp;gt; is a real number. We say that the sequence &amp;lt;math&amp;gt;(x_n)&amp;lt;/math&amp;gt; converges to a limit &amp;lt;math&amp;gt;L&amp;lt;/math&amp;gt; with [[order of convergence]] &amp;lt;math&amp;gt;q&amp;lt;/math&amp;gt; if the following holds:&lt;br /&gt;
&lt;br /&gt;
&amp;lt;math&amp;gt;\lim_{n \to \infty} \frac{|x_{n+1} - L|}{|x_n - L|^q} = \mu, \mu &amp;gt; 0&amp;lt;/math&amp;gt;&lt;br /&gt;
&lt;br /&gt;
In other words, if we were to try to construct a function that approximately predicted &amp;lt;math&amp;gt;|x_{n+1} - L|&amp;lt;/math&amp;gt; in terms of &amp;lt;math&amp;gt;|x_n - L|&amp;lt;/math&amp;gt;, the [[order of zero]] of a the function at zero would be &amp;lt;math&amp;gt;q&amp;lt;/math&amp;gt;. Note that convergence of this form is superlinear.&lt;br /&gt;
&lt;br /&gt;
Two special cases:&lt;br /&gt;
&lt;br /&gt;
* &amp;lt;math&amp;gt;q = 2&amp;lt;/math&amp;gt;, in which case we say the sequence &amp;#039;&amp;#039;&amp;#039;converges quadratically&amp;#039;&amp;#039;&amp;#039; or has &amp;#039;&amp;#039;&amp;#039;quadratic convergence&amp;#039;&amp;#039;&amp;#039;.&lt;br /&gt;
* &amp;lt;math&amp;gt;q = 3&amp;lt;/math&amp;gt;, in which case we say the sequence &amp;#039;&amp;#039;&amp;#039;converges cubically&amp;#039;&amp;#039;&amp;#039; or has &amp;#039;&amp;#039;&amp;#039;cubic convergence&amp;#039;&amp;#039;&amp;#039;.&lt;br /&gt;
&lt;br /&gt;
===Logarithmic convergence===&lt;br /&gt;
&lt;br /&gt;
{{further|[[logarithmic convergence]]}}&lt;br /&gt;
&lt;br /&gt;
If the sequence converges &amp;#039;&amp;#039;sublinearly&amp;#039;&amp;#039;, and further:&lt;br /&gt;
&lt;br /&gt;
&amp;lt;math&amp;gt;\lim_{n \to \infty} \frac{|x_{n+2} - x_{n+1}|}{|x_{n+1} - x_n|} = 1&amp;lt;/math&amp;gt;&lt;br /&gt;
&lt;br /&gt;
then we say that the convergence is &amp;#039;&amp;#039;logarithmic&amp;#039;&amp;#039;. This use of terminology differs somewhat from the others, because it is &amp;#039;&amp;#039;not&amp;#039;&amp;#039; the case that we can use a logarithmic function to approximate the description of &amp;lt;math&amp;gt;|x_{n+1} - L|&amp;lt;/math&amp;gt; in terms of &amp;lt;math&amp;gt;|x_n - L|&amp;lt;/math&amp;gt;. (So what&amp;#039;s the rationale of the term?)&lt;br /&gt;
&lt;br /&gt;
===Non-steady convergence or R-convergence===&lt;br /&gt;
&lt;br /&gt;
In some cases, the convergence of the sequence is not &amp;#039;&amp;#039;steadily&amp;#039;&amp;#039; of a certain type, but it still approximates that type.&lt;br /&gt;
&lt;br /&gt;
We say that a sequence &amp;lt;math&amp;gt;(x_n)&amp;lt;/math&amp;gt; has R-convergence of a particular type to a limit &amp;lt;math&amp;gt;L&amp;lt;/math&amp;gt; if we can find a sequence &amp;lt;math&amp;gt;(\varepsilon_n)&amp;lt;/math&amp;gt; such that, for all &amp;lt;math&amp;gt;n&amp;lt;/math&amp;gt;:&lt;br /&gt;
&lt;br /&gt;
&amp;lt;math&amp;gt;|x_n - L| &amp;lt; \varepsilon_n&amp;lt;/math&amp;gt;&lt;br /&gt;
&lt;br /&gt;
and the sequence &amp;lt;math&amp;gt;\varepsilon_n&amp;lt;/math&amp;gt; has convergence to 0 of that type.&lt;br /&gt;
&lt;br /&gt;
In particular, we can talk of R-linear convergence, R-superlinear convergence, R-sublinear convergence, R-quadratic convergence, R-logarithmic convergence, etc.&lt;/div&gt;</summary>
		<author><name>Vipul</name></author>
	</entry>
</feed>