Rate of convergence


Rate of convergence

In numerical analysis, the speed at which a convergent sequence approaches its limit is called the rate of convergence. Although strictly speaking, a limit does not give information about any finite first part of the sequence, this concept is of practical importance if we deal with a sequence of successive approximations for an iterative method, as then typically fewer iterations are needed to yield a useful approximation if the rate of convergence is higher. This may even make the difference between needing ten or a million iterations.

Similar concepts are used for discretization methods. The solution of the discretized problem converges to the solution of the continuous problem as the grid size goes to zero, and the speed of convergence is one of the factors of the efficiency of the method. However, the terminology in this case is different from the terminology for iterative methods.

Series acceleration is a collection of techniques for improving the rate of convergence of a series. Such acceleration is commonly accomplished with sequence transformations.

Contents

Convergence speed for iterative methods

Basic definition

Suppose that the sequence {xk} converges to the number L.

We say that this sequence converges linearly to L, if there exists a number μ ∈ (0, 1) such that

 \lim_{k\to \infty} \frac{|x_{k+1}-L|}{|x_k-L|} = \mu.

The number μ is called the rate of convergence.

If the sequences converges, and

  • μ = 0, then the sequence is said to converge superlinearly.
  • μ = 1, then the sequence is said to converges sublinearly.

If the sequences converges sublinearly and additionally

 \lim_{k\to \infty} \frac{|x_{k+2} - x_{k+1}|}{|x_{k+1} - x_k|} = 1,

then it is said the sequence {xk} converges logarithmically to L.

The next definition is used to distinguish superlinear rates of convergence. We say that the sequence converges with order q for q > 1 to L if

 \lim_{k\to \infty} \frac{|x_{k+1}-L|}{|x_k-L|^q} = \mu \,\big|\; \mu > 0.

In particular, convergence with order

  • 2 is called quadratic convergence,
  • 3 is called cubic convergence,
  • etc.

This is sometimes called Q-linear convergence, Q-quadratic convergence, etc., to distinguish it from the definition below. The Q stands for "quotient," because the definition uses the quotient between two successive terms.

Extended definition

The drawback of the above definitions is that these do not catch some sequences which still converge reasonably fast, but whose "speed" is variable, such as the sequence {bk} below. Therefore, the definition of rate of convergence is sometimes extended as follows.

Under the new definition, the sequence {xk} converges with at least order q if there exists a sequence {εk} such that

|x_k - L|\le\varepsilon_k^q\quad\mbox{for all }k,

and the sequence {εk} converges to zero with order q according to the above "simple" definition. To distinguish it from that definition, this is sometimes called R-linear convergence, R-quadratic convergence, etc. (with the R standing for "root").

Examples

Consider the following sequences:

\begin{align}
  a_0 &= 1 ,\,       &&a_1 = \frac12 ,\, &&a_2 = \frac14 ,\,    &&a_3 = \frac18 ,\,     &&a_4 = \frac1{16} ,\,    &&a_5 = \frac1{32} ,\, &&\ldots ,\, &&a_k = \frac1{2^k} ,\,                           &&\ldots \\
  b_0 &= 1 ,\,       &&b_1 = 1 ,\,       &&b_2 = \frac14 ,\,    &&b_3 = \frac14 ,\,     &&b_4 = \frac1{16} ,\,    &&b_5 = \frac1{16} ,\, &&\ldots ,\, &&b_k = \frac1{4^{\left\lfloor \frac{k}{2} \right\rfloor}} ,\, &&\ldots \\
  c_0 &= \frac12 ,\, &&c_1 = \frac14 ,\, &&c_2 = \frac1{16} ,\, &&c_3 = \frac1{256} ,\, &&c_4 = \frac1{65\,536} ,\,                      &&&&\ldots ,\, &&c_k = \frac1{2^{2^k}} ,\,                       &&\ldots \\
  d_0 &= 1 ,\,       &&d_1 = \frac12 ,\, &&d_2 = \frac13 ,\,    &&d_3 = \frac14 ,\,     &&d_4 = \frac15 ,\,       &&d_5 = \frac16 ,\,    &&\ldots ,\, &&d_k = \frac1{k+1} ,\,                           &&\ldots
\end{align}

The sequence {ak} converges linearly to 0 with rate 1/2. More generally, the sequence k converges linearly with rate μ if |μ| < 1. The sequence {bk} also converges linearly to 0 with rate 1/2 under the extended definition, but not under the simple definition. The sequence {ck} converges superlinearly. In fact, it is quadratically convergent. Finally, the sequence {dk} converges sublinearly.

Plot showing the different rates of convergence for the sequences a_k, b_k, c_k and d_k.
Linear, Linear, Superlinear/Quadratic and sublinear rate of convergence.

Convergence speed for discretization methods

A similar situation exists for discretization methods. Here, the important parameter is not the iteration number k but the number of grid points, here denoted n. In the simplest situation (a uniform one-dimensional grid), the number of grid points is inversely proportional to the grid spacing.

In this case, a sequence xn is said to converge to L with order p if there exists a constant C such that

| xnL | < Cn p for all n.

This is written as |xn - L| = O(n-p) using the big O notation.

This is the relevant definition when discussing methods for numerical quadrature or the solution of ordinary differential equations.

Examples

The sequence {dk} with dk = 1 / (k+1) was introduced above. This sequence converges with order 1 according to the convention for discretization methods.

The sequence {ak} with ak = 2-k, which was also introduced above, converges with order p for every number p. It is said to converge exponentially using the convention for discretization methods. However, it only converges linearly (that is, with order 1) using the convention for iterative methods.

Acceleration of convergence

Many methods exist to increase the rate of convergence of a given sequence, i.e. to transform a given sequence into one converging faster to the same limit. Such techniques are in general known as "series acceleration". The goal of the transformed sequence is to be much less "expensive" to calculate than the original sequence. One example of series acceleration is Aitken's delta-squared process.

References

The simple definition is used in

  • Michelle Schatzman (2002), Numerical analysis: a mathematical introduction, Clarendon Press, Oxford. ISBN 0-19-850279-6.

The extended definition is used in

  • Kendell A. Atkinson (1988), An introduction to numerical analysis (2nd ed.), John Wiley and Sons. ISBN 0-471-50023-2.
  • Walter Gautschi (1997), Numerical analysis: an introduction, Birkhäuser, Boston. ISBN 0-817-63895-4.
  • Endre Süli and David Mayers (2003), An introduction to numerical analysis, Cambridge University Press. ISBN 0-521-00794-1.

Logarithmic convergence is used in

The Big O definition is used in

  • Richard L. Burden and J. Douglas Faires (2001), Numerical Analysis (7th ed.), Brooks/Cole. ISBN 0-534-38216-9

The terms Q-linear and R-linear are used in; The Big O definition when using Taylor series is used in

  • Nocedal, Jorge; Wright, Stephen J. (2006). Numerical Optimization (2nd ed.). Berlin, New York: Springer-Verlag. pp. 619+620. ISBN 978-0-387-30303-1 .

One may also study the following paper for Q-linear and R-linear:

  • Potra, F. A. (1989). "On Q-order and R-order of convergence". J. Optim. Th. Appl. 63 (3): 415–431. doi:10.1007/BF00939805. 

Wikimedia Foundation. 2010.

Look at other dictionaries:

  • Convergence criteria — (also known as the Maastricht criteria) are the criteria for European Union member states to enter the third stage of European Economic and Monetary Union (EMU) and adopt the euro. The four main criteria are based on Article 121(1) of the… …   Wikipedia

  • Convergence (economics) — The idea of convergence in economics (also sometimes known as the catch up effect) is the hypothesis that poorer economies per capita incomes will tend to grow at faster rates than richer economies. As a result, all economies should eventually… …   Wikipedia

  • Rate function — In mathematics mdash; specifically, in large deviations theory mdash; a rate function is a function used to quantify the probabilities of rare events. It is required to have several nice properties which assist in the formulation of the large… …   Wikipedia

  • convergence and divergence — ▪ atmospheric       in meteorology, the accumulation or drawing apart of air, as well as the rate at which each takes place. The terms are usually used to refer specifically to the horizontal inflow (convergence) or outflow (divergence) of air.… …   Universalium

  • Radius of convergence — In mathematics, the radius of convergence of a power series is a quantity, either a non negative real number or ∞, that represents a domain (within the radius) in which the series will converge. Within the radius of convergence, a power series… …   Wikipedia

  • Contraction and Convergence — (C C) is a proposed global framework for reducing greenhouse gas emissions to combat climate change. Conceived by the Global Commons Institute [GCI] in the early 1990s, the Contraction and Convergence strategy consists of reducing overall… …   Wikipedia

  • Euro convergence criteria — Convergence criteria redirects here. For the mathematical term, see Series (mathematics)#Convergence criteria. Euro adoption by EU member states v · d · e …   Wikipedia

  • Technological convergence — This article is about technology convergence including convergence of media technology. For consolidation of media ownership, see Concentration of media ownership. Technological convergence is the tendency for different technological systems to… …   Wikipedia

  • European Exchange Rate Mechanism — The European Exchange Rate Mechanism, ERM, was a system introduced by the European Community in March 1979, as part of the European Monetary System (EMS), to reduce exchange rate variability and achieve monetary stability in Europe, in… …   Wikipedia

  • Newton's method — In numerical analysis, Newton s method (also known as the Newton–Raphson method), named after Isaac Newton and Joseph Raphson, is a method for finding successively better approximations to the roots (or zeroes) of a real valued function. The… …   Wikipedia