Download presentation
Presentation is loading. Please wait.
1
Fixed- Point Iteration
Sec:2.2 Fixed- Point Iteration
2
Sec:2.2 Fixed- Point Iteration
The number ๐ is a fixed point for a given function ๐ if. ๐=๐ ๐ Example: A fixed point for g occurs precisely when the graph of ๐ฆ = ๐(๐ฅ) intersects the graph of ๐ฆ = ๐ฅ Determine any fixed points of the function ๐(๐ฅ) = ๐ฅ 2 โ 2. solution ๐=๐(๐) ๐= ๐ 2 โ 2 0= ๐ 2 โ๐โ 2 ๐=โ1, ๐=2
3
|gโ(x)| โค k, for all x โ (a, b),
Sec:2.2 Fixed- Point Iteration Theorem 2.3 (i) If g โ C[a, b] and g(x) โ [a, b] for all x โ [a, b], then g has at least one fixed point in [a, b]. (ii) If, in addition, g(x) exists on (a, b) and a positive constant k < 1 exists with |gโ(x)| โค k, for all x โ (a, b), then there is exactly one fixed point in [a, b]. Example Show that g(x) = ๐ฅ 2 โ1 3 has a unique fixed point on the interval [โ1, 1]. Example Show that g(x) = 3 โ๐ฅ has at least one fixed point on the interval [0, 1].
4
Sec:2.2 Fixed- Point Iteration
The number ๐ is a fixed point for a given function ๐ if.๐=๐ ๐ Example ๐ ๐ ๐ Find the fixed point of the function ๐(๐)= ๐ โ๐ ย ๐ ๐+๐ =๐( ๐ ๐ ) ๐ ๐+๐ = ๐ โ ๐ ๐ ย Starting with an initial guess of p0 = 0 ๐ ๐ =๐ ๐ ๐ = ๐ โ๐ =๐ ๐ ๐ =๐ ๐ ๐ = ๐ โ๐ = ๐ ๐ =๐ ๐ ๐ = ๐ โ๐.๐๐๐๐๐๐ = Thus, each iteration brings the estimate closer to the true fixed point: โฎ โฎ โฎ โฎ
5
Sec:2.2 Fixed- Point Iteration
The number ๐ is a fixed point for a given function ๐ if. ๐=๐ ๐ Finding fixed point of ๐(๐) Root Finding of ๐(๐) ๐ Is a fixed point of ๐(๐) ๐ Is a root of ๐(๐) ๐ (๐) = ๐ โ ๐(๐)
6
Sec:2.2 Fixed- Point Iteration
Example1 ๐ฅ 2 โ 2๐ฅ + 3 = 0 can be simply manipulated to yield simple fixed-point iteration Step 1 ย ๐ฅ= ๐ฅ ๐ฅ=๐(๐ฅ) rearranging the function ๐ (๐) = ๐ so that x is on the left-hand side of the equation: ๐ = ๐(๐) (*) Example2 sinโก(๐ฅ)= 0 can be simply manipulated to yield ๐ฅ=๐ฅ+sinโก(๐ฅ) ๐ฅ=๐(๐ฅ) Step 2 given an initial guess at the root ๐ ๐ , (*) can be used to compute a new estimate ๐ ๐+1 as expressed by the iterative formula ๐ ๐+๐ =๐( ๐ ๐ ) Note Convert the problem from root-finding to finding fixed-point
7
Sec:6.1 Fixed- Point Iteration
Step 1 Example1 Use simple fixed-point iteration to locate the root of ๐ ๐ = ๐ โ๐ โ๐ rearranging the function ๐= ๐ โ๐ ย ๐=๐(๐)ย ๐ ๐ Step 2 ๐ ๐ ๐+๐ =๐( ๐ ๐ ) ๐ ๐+๐ = ๐ โ ๐ ๐ ย Starting with an initial guess of p0 = 0 ๐ ๐ =๐ ๐ ๐ = ๐ โ๐ =๐ ๐ ๐ =๐ ๐ ๐ = ๐ โ๐ = ๐ ๐ =๐ ๐ ๐ = ๐ โ๐.๐๐๐๐๐๐ = โฎ โฎ โฎ โฎ Thus, each iteration brings the estimate closer to the true value of the root:
8
Sec:2.2 Fixed- Point Iteration
๐บ ๐ = ๐ ๐ โ๐ ๐ ร๐๐๐% clear; clc; format long p(1) = 0; g exp(-x); true_root = ; for k=1:11 p(k+1) = g( p(k) ); rel=abs( (p(k) - true_root )/ true_root ); fprintf('%d %10.4f %10.4f\n', k,p(k),rel); end ๐ ๐ ๐บ ๐ (%) ๐ Notice that the true percent relative error for each iteration is roughly proportional (by a factor of about 0.5 to 0.6) to the error from the previous iteration. This property, called linear convergence
9
Sec:2.2 Fixed- Point Iteration
The derivative mean-value theorem states that if a function ๐(๐ฅ) and its first derivative are continuous over an interval ๐ โค ๐ฅ โค ๐, then there exists at least one value of ๐ฅ= ๐ within the interval such that ๐โฒ(๐ ) = ๐(๐) โ ๐(๐) ๐ โ ๐ The right-hand side of this equation is the slope of the line joining ๐(๐) and ๐(๐). ๐ ๐ โ ๐ ๐ = ๐ โฒ ๐ (๐โ๐) Slope of the tangent line is ๐โฒ(๐ ) Slope of this line is ๐(๐) โ ๐(๐) ๐ โ ๐ ๐
10
๐ฌ ๐๐ <๐ ๐ฌ ๐ <โฏ< ๐ ๐ ๐ฌ ๐ < ๐ ๐๐ ๐ฌ ๐
Convergence Sec:2.2 Fixed- Point Iteration Suppose that the true solution is ๐ฌ ๐+๐ = ๐โฒ ๐ โ ๐ฌ ๐ ๐=๐(๐)ย <๐ ๐ฌ ๐ the iterative equation is ( If ๐ ๐ โค๐, ๐<๐ ) ๐ ๐+๐ =๐( ๐ ๐ )ย the errors decrease with each iteration Subtracting these equations yields ๐ฌ ๐+๐ <๐ ๐ฌ ๐ ๐โ ๐ ๐+๐ =๐ ๐ โ๐( ๐ ๐ )ย ๐ฌ ๐๐ <๐ ๐ฌ ๐ <โฏ< ๐ ๐ ๐ฌ ๐ < ๐ ๐๐ ๐ฌ ๐ The derivative mean-value theorem gives ๐ฌ ๐ < ๐ ๐ ๐ฌ ๐ ๐โ ๐ ๐+๐ =๐โฒ ๐ (ย ๐โ ๐ ๐ ) Step 1 where ๐ is somewhere between ๐ and ๐ ๐ rearranging the function ๐ (๐) = ๐ ย so that x is on the left-hand side of the equation: ๐ = ๐(๐) If the true error for iteration n is defined as ๐ฌ ๐ =๐โ ๐ ๐ ๐ฌ ๐+๐ =๐โฒ ๐ ๐ฌ ๐ Select ๐ ๐ ๐๐ ๐๐๐๐ ๐โฒ ๐ <๐ ๐ฌ ๐+๐ = ๐โฒ ๐ โ ๐ฌ ๐
11
Sec:2.2 Fixed- Point Iteration
Theorem 2.4 (Fixed-Point Theorem) Let g โ C[a, b] be such that g(x) โ [a, b], for all x in [a, b]. Suppose, in addition, that gโ exists on (a, b) and that a constant 0 < k < 1 exists with |g(x)| โค k, for all x โ (a, b). Then for any number p0 in [a, b], the sequence defined by pn = g( pnโ1), n โฅ 1, converges to the unique fixed point p in [a, b]. Corollary 2.5 If g satisfies the hypotheses of Theorem 2.4, then bounds for the error involved in using pn to approximate p are given by | pn โ p| โค ๐ ๐ max{ p0 โ a, b โ p0} (2.5) and | pn โ p| โค ๐ ๐ ๐โ๐ | p1 โ p0|, for all n โฅ 1. (2.6) Two important error equation
12
Sec:2.2 Fixed- Point Iteration
Corollary 2.5 If g satisfies the hypotheses of Theorem 2.4, then bounds for the error involved in using pn to approximate p are given by | pn โ p| โค ๐ ๐ max{ p0 โ a, b โ p0} (2.5) and | pn โ p| โค ๐ ๐ ๐โ๐ | p1 โ p0|, for all n โฅ 1. (2.6) Remark 1 The smaller the value of k, the faster the convergence, which may be very slow if k is close to 1, Remark 2 the rate at which { pn } converges
13
Sec:2.2 Fixed- Point Iteration
The function g(x) is decreasing ๐ 0.2 =0.8187 0.407, โ[0.2, 0.9] ๐ 0.9 =0.407 g(x) โ [a, b], for all x in [a, b]
14
Sec:2.2 Fixed- Point Iteration
| pn โ p| โค ๐ ๐ ๐โ๐ | p1 โ p0|, for all n โฅ 1. (2.6) | pn โ p| โค ๐ ๐ max{ p0 โ a, b โ p0} (2.5) | pn โ p| โค ( ๐ โ๐.๐ ) ๐ max{ 0.8 โ 0.2, 0.9 โ 0.8} | pn โ p| โค ( ๐ โ0.2 ) ๐ 1โ ๐ โ โ0.8 | pn โ p| โค ( ๐ โ๐.๐ ) ๐ { 0.6 } | pn โ p| โค ( ๐ โ0.2 ) ๐ 1โ ๐ โ0.2 { } โค 10 โ2 | pn โ p| โค ( ๐ โ๐.๐ ) ๐ { ๐.๐ } โค ๐๐ โ๐ ๐ โ0.2 ๐ โค 10 โ2 (1โ ๐ โ0.2 ) ๐ โ๐.๐ ๐ โค ๐๐ โ๐ ๐.๐ โ0.2 ๐โค๐๐( 10 โ2 (1โ ๐ โ0.2 ) ) โ ๐โฅ โ๐.๐ ๐โค๐๐( ๐๐ โ๐ ๐.๐ ) โ ๐โฅ๐๐.๐๐๐๐ โ ๐=27 โ ๐=๐๐
15
Sec:2.2 Fixed- Point Iteration
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.