Presentation is loading. Please wait.

Presentation is loading. Please wait.

Newton’s Method Other Recursive Methods Modified Fixed Point Method.

Similar presentations


Presentation on theme: "Newton’s Method Other Recursive Methods Modified Fixed Point Method."— Presentation transcript:

1 Newton’s Method Other Recursive Methods Modified Fixed Point Method

2 Tangential Roots The consistent problem we have encountered with both the Bisection Method and the Regula-Falsi Method is that of a tangential root. In other words to apply the method the function f(x) must be of opposite sign on each side of the root (i.e. f(a)  f(b) <0). We want to be able to address this problem and still find a root if the function f(x) crosses the x -axis. Modified Fixed Point Algorithm The fixed point algorithm can be modified to find roots instead of fixed points. This comes from a simple algebra fact that if w 0 is a fixed point of h(x) and h(x)=f(x)+x then w 0 will be a root of f(x). h(w 0 ) = w 0 ( x 0 is a fixed point of h(x) ) h(w 0 ) = f(w 0 ) + w 0 (definition of h(x) ) f(w 0 ) + w 0 = w 0 (substitute) f(w 0 ) =0

3 Below we show how the modified fixed point algorithm can be applied to find roots. We do this for a couple of functions. nxnxn h(x n ) 017 171.85714 2 4.23077 3 2.41818 nxnxn h(x n ) 017 17-29 2 -893 3 -799229

4 The problem with the modified fixed point method is that it is very computationally unstable for many functions with certain initial values. By this I mean that the sequence that it generates does not converge. Notice that in the previous example even though we were trying to compute the same root in both examples (i.e. 3) in the first case when dividing by x the algorithm converged. Newton’s Method The idea for this method is to use f(x) to build another function h(x) that will generate a recursive sequence that converges to the root just like the modified fixed point method. The idea here is to keep following the tangent line at a point on the graph down to the x -axis and use that for the value of x that will approximate the root. In other words h(x) represents the x -intercept of the tangent line of f(x). x f(x) h(x) f(x n ) h(x n )=x n+1 xnxn root

5 To get what x n is from x n+1 we write the equation of the tangent line at x n, plug in the point ( x n+1,0) and solve for x n+1. equation of tangent at x n substitute in ( x n+1,0 ) solve this is the h(x) The equation above gives the recursively defined sequence for x n. This is what is used for Newton’s Method. The halting condition is usually given by the Standard Cauchy Error.

6 nxnxn h(x n ) 011.5 1 1.41667 2 1.41422 3 1.41421 nxnxn h(x n ) 015 153.46575 2 2.53422 3 2.059 Here are some examples of Newton’s Method applied.

7 Problems With Newton’s Method The number of iterations required can not be determined before the algorithm begins. The algorithm will not work if f(x) is not differentiable. The algorithm will halt (program termination by division by zero if not checked for) if a horizontal tangent line is encountered. Newton’s method will sometimes find an extraneous root.


Download ppt "Newton’s Method Other Recursive Methods Modified Fixed Point Method."

Similar presentations


Ads by Google