Approximate Integration: Discussion Exercises for Math 1B, Assignments of Calculus

Discussion exercises for math 1b students, focusing on approximation techniques for integrals. The exercises cover various methods, including midpoint, trapezoid, and simpson rules, and include examples and hints. Students are encouraged to work in groups and discuss their solutions.

Typology: Assignments

Pre 2010

Uploaded on 10/01/2009

koofers-user-n2a
koofers-user-n2a 🇺🇸

10 documents

1 / 2

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
Math 1B: Discussion Exercises
GSI: Theo Johnson-Freyd
http://math.berkeley.edu/~theojf/09Summer1B/
Find two or three classmates and a few feet of chalkboard. Introduce yourself to your new
friends, and write all of your names at the top of the chalkboard. As a group, try your hand at
the following exercises. Be sure to discuss how to solve the exercises how you get the solution
is much more important than whether you get the solution. If as a group you agree that you all
understand a certain type of exercise, move on to later problems. You are not expected to solve all
the exercises: some are very hard.
Many of the exercises are from Single Variable Calculus: Early Transcendentals for UC Berkeley
by James Stewart; these are marked with an §. Others are my own, are from the mathematical
folklore, or are independently marked.
Here’s a hint: drawing pictures e.g. sketching graphs of functions will always make the
problem easier.
Approximate Integration
To approximate Rb
af(x)dx, let x= (ba)/n,xi=a+ix, and ¯xi= (xi1+xi)/2. Then
define the following approximations (we use different number from Stewart for Simpson’s Rule Sn):
Ln= (f(x0) + · ·· +f(xn1)) x Rn= (f(x1) + · ·· +f(xn)) x
Tn= (f(x0)+2f(x1) + · ·· + 2f(xn1) + f(xn)) x
2Mn= (fx1) + · ·· +fxn)) x
Sn= (f(x0)+4fx1)+2f(x1)+4fx2)+2f(x2) + · ·· + 2f(xn1)+4fxn) + f(xn)) x
6
These have the following errors:
|EL| sup
x[a,b]
f0(x)
(ba)2
2n|ER| sup
x[a,b]
f0(x)
(ba)2
2n
|ET| sup
x[a,b]
f00(x)
(ba)3
12n2|EM| sup
x[a,b]
f00(x)
(ba)3
24n2
|ES| sup
x[a,b]
f(4)(x)
(ba)5
2780n4
The word “sup” is short for “supremum” the symbol “supx[a,b]g(x)” means “the largest value
of g(x) as xranges over [a, b]”. In practice, it suffices to replace the suprema with some easy-to-
compute numbers which are even bigger.
1. Explain each of the above approximation techniques when n= 1.
2. Let f(x) be a positive increasing function with negative second derivative on [a, b]. Place the
following five numbers in increasing order: Ln,Rn,Tn,Mn, and Rb
af(x)dx.
3. If you were to evaluate each of the following integrals using each of the Midpoint, Trapezoid,
and Simpson Rules with 5 subintervals, what would be your expected error? How many
1
pf2

Partial preview of the text

Download Approximate Integration: Discussion Exercises for Math 1B and more Assignments Calculus in PDF only on Docsity!

Math 1B: Discussion Exercises

GSI: Theo Johnson-Freyd http://math.berkeley.edu/~theojf/09Summer1B/

Find two or three classmates and a few feet of chalkboard. Introduce yourself to your new friends, and write all of your names at the top of the chalkboard. As a group, try your hand at the following exercises. Be sure to discuss how to solve the exercises — how you get the solution is much more important than whether you get the solution. If as a group you agree that you all understand a certain type of exercise, move on to later problems. You are not expected to solve all the exercises: some are very hard. Many of the exercises are from Single Variable Calculus: Early Transcendentals for UC Berkeley by James Stewart; these are marked with an §. Others are my own, are from the mathematical folklore, or are independently marked. Here’s a hint: drawing pictures — e.g. sketching graphs of functions — will always make the problem easier.

Approximate Integration

To approximate

∫ (^) b a f^ (x)^ dx, let ∆x^ = (b^ −^ a)/n,^ xi^ =^ a^ +^ i∆x, and ¯xi^ = (xi−^1 +^ xi)/2. Then define the following approximations (we use different number from Stewart for Simpson’s Rule Sn):

Ln = (f (x 0 ) + · · · + f (xn− 1 )) ∆x Rn = (f (x 1 ) + · · · + f (xn)) ∆x

Tn = (f (x 0 ) + 2f (x 1 ) + · · · + 2f (xn− 1 ) + f (xn))

∆x 2

Mn = (f (¯x 1 ) + · · · + f (¯xn)) ∆x

Sn = (f (x 0 ) + 4f (¯x 1 ) + 2f (x 1 ) + 4f (¯x 2 ) + 2f (x 2 ) + · · · + 2f (xn− 1 ) + 4f (¯xn) + f (xn))

∆x 6

These have the following errors:

|EL| ≤ sup x∈[a,b]

∣f ′(x)

∣ (b^ −^ a)

2 2 n

|ER| ≤ sup x∈[a,b]

∣f ′(x)

∣ (b^ −^ a)

2 2 n

|ET | ≤ sup x∈[a,b]

f ′′(x)

∣∣ (b − a)^3 12 n^2

|EM | ≤ sup x∈[a,b]

f ′′(x)

∣∣ (b − a)^3 24 n^2

|ES | ≤ sup x∈[a,b]

∣∣f (4)(x)

∣∣ (b^ −^ a)^5 2780 n^4

The word “sup” is short for “supremum” — the symbol “supx∈[a,b] g(x)” means “the largest value of g(x) as x ranges over [a, b]”. In practice, it suffices to replace the suprema with some easy-to- compute numbers which are even bigger.

  1. Explain each of the above approximation techniques when n = 1.
  2. Let f (x) be a positive increasing function with negative second derivative on [a, b]. Place the following five numbers in increasing order: Ln, Rn, Tn, Mn, and

∫ (^) b a f^ (x)dx.

  1. If you were to evaluate each of the following integrals using each of the Midpoint, Trapezoid, and Simpson Rules with 5 subintervals, what would be your expected error? How many

subintervals would you need to ensure an error less than 0.00001?

(a)

0

x dx (b)

0

z e−z^ dz (c)

4

ln(x^3 + 2) dx

  1. (a) Show that (Ln + Rn)/2 = Tn.

(b) § Show that (Tn + Mn)/2 = T 2 n. (c) § Show that (Tn + 2Mn)/3 = Sn.

  1. By explicit calculation, show that Simpson’s rule calculates the area under a cubic curve exactly. What are the highest degree polynomials the rest of the approximation rules calculate exactly?
  2. By explicit calculation, show that the errors for Ln and Rn are exact when f (x) is a linear function, and that the errors for Tn and Mn are exact when f (x) is a quadratic function.
  3. Make sense of the following proof from Proofs without Words: Exercises in Visual Thinking by Roger B. Nelsen (1993):