472 Assignment 6

ENSC 180 Introduction to Engineering Analysis Tools
Assignment #6: Numerical Integration
Instructions
The assignment is due: Tuesday, April 14, 2015, by 11:50pm
Your submission must include the following files:

integralSimpsons.m
adaptiveSimpsons.m
demoSimpsons.m
integralTrap.m
adaptiveTrap.m
demoTrap.m
f.p (provided as supplementary file)
report.pdf
Please note that the scripts contained in demoSimpsons.m and demoTrap.m must run
without error, otherwise your code will be given a grade of 0.
The report is an essential portion of this assignment. Please address each of the 3 points
given in the Analysis section of the assignment instructions (last page of the assignment). Be
concise; the purpose of the report is to address these 3 points and nothing else.
Any file other than the ones listed above is not necessary for submission.
Any MATLAB code used in your analysis can be copied directly into the report.
Background
The biggest drawback of composite quadrature rules is the dilemma of the step-size that will be
used. For example, if we have a monotonic function that increases at a steady rate, we may be
able to use a large step-size to obtain a highly accurate estimate of its integral while minimizing
computing time. However, this large step-size may fail miserably when applied to a wildly
varying function. In this case, we would prefer a smaller step-size. An even greater dilemma
occurs if we are faced with a function with both of these characteristics what would we do
then? We must look towards an adaptive approach where the function is recursively broken
down into smaller and smaller intervals, but only where necessary. An interval will be broken
down into two subintervals only if the integral done independently on the two subintervals is far
more accurate. This is known as adaptive quadrature.
Objectives
This assignment will cover the following topics:
1. Adaptive quadrature methods for numerical integration.
2.
3.
4.
5.
Recursive programming in MATLAB.

Line plotting tools to help visualize results.
Passing functions as arguments using function handles.
Simple computational complexity analysis using tic and toc functions.
You are required to integrate an unknown function across a finite range using adaptive
quadrature. You will gain insight on how the adaptive algorithm works by creating visualizations
of its optimally-chosen intervals of integration.
Preliminary: The unknown function that you are given is positive within an interval [1, 9] and 0
elsewhere. Before moving on, it would be wise to have a look at it. It is a function of only one
variable (i.e., of the form
) and can be sampled at any point. This unknown function is
defined in f.m.
Task 1: Write a MATLAB function with the following description:
function estIntegral = integralSimpsons(f, a, b)
This function should be stored in a file named integralSimpsons.m. It should estimate integral
of a function across a specified interval using the Simpsons rule, where:
f a MATLAB function handle.
a scalar value defining the lower limit of integration.
b scalar value defining the upper limit of integration.
estIntegral the value of the estimated integral.
Task 2: Write a MATLAB function with the following description:

function [estIntegral, intervals] = adaptiveSimpsons(f, a, b, tol, s)
This function will be called recursively to perform the numerical integration. It will also return
the endpoints of the M subintervals formed throughout the process in an M-by-2 matrix
intervals. This function should be stored in a file named adaptiveSimpsons.m and it
should perform the following steps:
1. Find the midpoint of the current interval [a, b] and store the result in a variable c.
2. Apply Simpsons rule, using integralSimpsons twice on the unknown function
(pointed at by MATLAB function handle f): Once on the left subinterval [a, c], and
again on the right subinterval [c, b].
3. In the current recursion, we must check for the stopping criteria. Let
denote the
estimate of our integral across an interval
. If
where
is tol in our implementation, stop the recursion and:
a. calculate the sum of
, and the following correction term
Store the result in estIntegral.

b. Store the subintervals [a, c] and [c, b] as rows a matrix intervals.
c. Exit out of the function using the return keyword.
4. If the stopping criteria is not satisfied, continue on with the recursion as follows:
a. Call the adaptiveSimpsons function twice: once on the left subinterval and
again on the right subinterval. Use half of the current tol as the error tolerance.
Store the outputs of both function calls in appropriately named variables.
b. Calculate the sum of the estimated integral outputs for these two recursive
function calls and store the result in estIntegral.
c. Vertically concatenate the matrix outputs containing the intervals created by these
two recursive function calls and store the result in intervals. You can do this
either by using matrix construction syntax, or by using the MATLAB cat
function.
d. Exit out of the function using the return keyword.
Task 3: Write a script to make it easier to execute these functions properly, and to create the
visualizations for the intervals. Give the script an intuitive name like demoSimpsons.m and
organize it as follows:
1. It is wise to start any script with:
clear all
close all
to ensure that it is entirely self-contained and can run with an initially empty workspace.
2. Ensure that the unknown function defined in f.m is inside your working folder along
with functions from Task 1 and 2. A MATLAB function handle is a data type that points
to a function and allows it to be passed as an argument of another function. For example,
minHandle = @(x) min(x) + max(x); would add the minimum and maximum
values of an input x when called as minHandle(x). Create a function handle that can
be used to call our unknown function f.
3. We wish to integrate the unknown function only in the interval [1, 9], where it is not 0.
Store the upper and lower limits in appropriate variables.
4. Set the tolerance, tol = 1e-2.

5. Obtain an initial estimate of the integral using integralSimpsons.
6. Call the function adaptiveSimpsons with this initial estimate and the previously
defined parameters.
7. Plot the unknown function using the plot command and a generous number of sample
points.
8. Query the y-axis limits using the get command on the current axis. To see which
attribute contains the y-axis limits, you can create a dummy plot,
e.g., plot(x, x.^2), select the axis object using the arrow, open up the Property
Editor under View of the figure window, and click on More Properties.
9. Set hold on and iterate through the rows of the intervals matrix.
a. Using the MATLAB line function, draw a vertical line that begins at the lower
y-axis limit and terminates at the upper y-axis limit for each x-axis value where an
interval endpoint occurs. Assign a variable to store the handle of this line object
for the next step, i.e., h = line(). (Hint: a line object is actually what is
created whenever you use the function plot, except the plot function also
makes a figure and axis to go along with it. So all we are doing is overlaying
straight lines on top of our original plot).
b. Use the set command with the line handle h to make the line dotted (':') and
black ([0, 0, 0]). Find the appropriate properties by searching lineseries
properties in the documentation.
10. Set hold off and try running everything.
Task 4: Once the previous tasks have been completed, the next step is to repeat them using the
trapezoid rule. Make copies of the files integralSimpsons.m, adaptiveSimpsons.m, and
demoSimpsons.m. Rename the copies appropriately: integralTrap.m, adaptiveTrap.m, and
demoTrap.m. Make sure the function names within the files are consistent with the new
filenames. There are only two meaningful changes here:
1) Inside integralTrap.m, you must repeat Task 1 for the trapezoid rule.
2) Inside adaptiveTrap.m, there is a very slight change in the stopping criteria:
and the correction term becomes:
All other changes are file/function name-related and are trivial.
Analysis
1. Compare the results obtained from adaptive Simpsons rule with those from the adaptive
trapezoid rule. In either case, why are the intervals chosen the way they are by the
algorithm? You may wish to vary the tolerances and re-run both cases to obtain
presentable result.
2. Compare the estimates for various tolerances with the MATLAB integral function,
which numerically evaluates the integral for a given function handle on a specified
interval. To get MATLAB to show more digits in your answer, type format long in
the terminal. This change may be reverted by typing format short, or simply
format, in the terminal.
3. Computational complexity is usually determined by the number of multiplications and
divisions. Since our function is unknown, we cannot accurately measure this. An
alternative is to measure the amount of time it takes to run our implementations
integrateSimpsons and integrateTrap on the unknown function. Use the
MATLAB tic and toc functions to perform the necessary measurements. Perform the
measurements many times and take the average to make sure they are reliable. Which
method is more computationally demanding? Why? Which method is more
computationally demanding in the context of the adaptive algorithm? Why?

472 Assignment 6

Uploaded by

Copyright:

Available Formats

472 Assignment 6

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

472 Assignment 6

Uploaded by

Copyright:

Available Formats

ENSC 180 Introduction to Engineering Analysis Tools

Assignment #6: Numerical Integration

The assignment is due: Tuesday, April 14, 2015, by 11:50pm

Your submission must include the following files:

Recursive programming in MATLAB.

Task 2: Write a MATLAB function with the following description:

is tol in our implementation, stop the recursion and:

a. calculate the sum of

, and the following correction term

Store the result in estIntegral.

4. Set the tolerance, tol = 1e-2.

and the correction term becomes:

All other changes are file/function name-related and are trivial.

You might also like