Learning Objectives

At the end of this class you should be able to

Reading

The textbook discusses root-finding in two separate contexts. Newton's method is used as an example of a computational algorithm Chapter 2, section 2.7.1 and then again in Chapter 7, section 7.1 through 7.3 as a practical means of discussing function m-files. Please review those sections.

Root-finding as a generic numerical method is discussed Chapter 14, section 14.1.

Download

Review of MATLAB Programming Techniques

As of last week, we have covered the basic MATLAB programming techniques that we need for the rest of this class. Of course, continued practice with those techniques is necessary and useful. However, there are only a few relatively specialized skills that we will need to add to your repertoire.

Here is a list of topics to review.

Please consult the reference pages of the class web site, and in particular the MATLAB Vocabulary reference document

Shifting Emphasis to Numerical Methods.

Having covered the basic programming techniques used in MATLAB, we now begin applying MATLAB to a basic problem in numerical methods. The first technique is root-finding. Since root-finding involves iteration, we also need to understand how to check for convergence of a sequence. This task requires combining knowledge of how to implement loops, if constructs, along with a little knowledge of floating point arithmetic.

Convergence of the Series Approximations to

In the Lab Exercise 03 the series approximation to was used to demonstrate a for loop.

We want to avoid the simplistic implementation by using a termination criterion to automatically stop evaluating terms in the series. This leads us to consider the difference between an absolute criterion and a relative criterion.

Convergence of Newton's Method for Computing .

In an application of a while loop in Lab Exercise 04, we used Newton's method to find the square root of a number. This also leads us to consider the difference between an absolute convergence criterion and a relative convergence criterion