

Study with the several resources on Docsity
Earn points by helping other students or get them with a premium plan
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
Community
Ask the community for help and clear up your study doubts
Discover the best universities in your country according to Docsity users
Free resources
Download our free guides on studying techniques, anxiety management strategies, and thesis advice from Docsity tutors
Solutions to worksheet 4 of math 143-03 at st. Louis university, focusing on the convergence of improper integrals using the comparison theorem and various approximation methods such as trapezoid rule, midpoint rule, and simpson's rule.
Typology: Assignments
1 / 3
This page cannot be seen from the preview
Don't miss anything!
St. Louis University
Spring 2008
Math 143–
Prof. G. Marks
∞
1
ln x
1 + x 2
dx convergent or divergent? Substitute u = 1/x in this integral to equate it with a different
improper integral. Then determine
0
ln x
1 + x 2
dx. Illustrate your answer with a graph.
Solution. To avoid the difficulty of integrating
ln x
1 + x^2
, we use the Comparison Theorem. For all x in [1, ∞),
ln x
1 + x 2
ln x
x 2
furthermore,
∗
∞
1
ln x
x 2
dx =
1 + ln x
x
]x=∞
x=
= 1 − lim x→∞
1 + ln x
x
L′^ H = 1 − lim x→∞
x
1
converges. So
∞
1
ln x
1 + x 2
dx must also converge, by the Comparison Theorem.
†
Now substitute u = 1/x to obtain
∫ (^) x=∞
x=
ln x
1 + x 2
dx =
∫ (^) u=
u=
ln (1/u)
1 + (1/u)
2
u 2
du =
∫ (^) u=
u=
ln u
u 2
du = −
∫ (^) u=
u=
ln u
u 2
du.
This last integral is equal to −
∫ (^) x=
x=
ln x
x 2
dx. (We’ve just changed the dummy variable from “u ” to “ x.”) Hence,
x=∞
x=
ln x
1 + x 2
dx =
x=
x=
ln x
1 + x 2
dx +
x=∞
x=
ln x
1 + x 2
dx = 0.
(We just showed that these last two integrals are each other’s negatives, and that each converges, i.e. equals a finite
quantity. Therefore they cancel to give 0.) Graphing the integrand is left to you.
1
0
1 + x 3 dx.
(a) Write a formula for the Trapezoid Rule approximation with 100 subdivisions. Find a bound on the error. How
many subdivisions would suffice to compute the integral to within ± 10 − 12 ?
∗ Here and subsequently, the symbol
L′^ H = will be used to mean “equals, by L’Hospital’s Rule.” One must note in each case that the
conditions of L’Hospital’s Rule are satisfied. † This argument illustrates the proper explanation to give when proving convergence of an improper integral using the Comparison
Theorem. Statements like
“for large x, the 1 becomes insignificant compared with the x 2 , so
ln x
1 + x^2
≈
ln x
x^2
”
and
“the 1 becomes insignificant, so as far as convergence is concerned,
Z (^) ∞
1
ln x
1 + x 2 dx behaves like
Z (^) ∞
1
ln x
x 2 dx ”
do not belong in a mathematical argument. Mathematics is based on precise, provable assertions, not vague, subjective claims about
something being much larger than something else.
That having been said, it should be added that the intuitive mental process, which enables you to guess with some degree of accuracy
whether the integral converges, does involve thoughts such as: “If we replace 1 + x 2 by x 2 in the denominator, it should not affect
convergence.” Such thoughts guide us toward formulating a correct solution, but they are too imprecise to belong in a solution.
Hughes-Hallett provides dubious guidance in this matter. In Examples 1 and 3 on pp. 334–335, each solution begins with a vague,
intuitive argument, which admirably illustrates how one might think up a solution, but which does not belong in the solution itself.
The solution to Example 1 should not have included anything prior to: “[W]e observe that for 0 6 x 3 6 x 3
solution to Example 3 should not have included anything prior to: “We know that (ln t) − 1 < ln t < t for all positive t. ” (Actually,
this statement requires some justification. Note that if we define the function g(t) = t − ln t with domain t > 0 , then the derivative
g ′ (t) = 1 − 1 /t is negative whenever 0 < t < 1 and positive whenever t > 1; therefore, g(1) = 1 is the absolute minimum of g. Hence
g(t) = t − ln t > 1 ⇒ t > 1 + ln t for all t > 0. )
(b) Write a formula for the Midpoint Rule approximation with 100 subdivisions. Find a bound on the error. How
many subdivisions would suffice to compute the integral to within ± 10 − 12 ?
(c) Write a formula for the Simpson’s Rule approximation with 100 subdivisions. Find a bound on the error.
How many subdivisions would suffice to compute the integral to within ± 10 − 12 ?
Solution. Let f (x) =
1 + x 3
. If 0 6 x 6 1 ,
|f
′′ (x)| =
3 x
4
4 (1 + x 3 )
3 / 2
3 x
4
3 x
4
12 x
and
|f
′′′′ (x)| =
9 x
8
5 − 720 x
2
16 (1 + x 3 )
7 / 2
9 x
8
5 − 720 x
2
9 x
8
504 x
5
− 720 x
2
Thus, the respective errors ET , EM , and ES in the Trapezoid Rule approximation, the Midpoint Rule approxi-
mation, and the Simpson’s Rule approximation satisfy
(b − a) 3
12 n 2
f
′′ (c)
12 n 2
16 n 2
(b − a) 3
24 n 2
f
′′ (z)
24 n 2
32 n 2
(b − a)
5
180 n^4
f
′′′′ (ξ)
180 n^4
320 n^4
(where c, z, and ξ are unknown points in the interval [a, b] = [0, 1], and n is the number of subdivisions).
(a)
1
0
1 + x 3 dx ≈
99 ∑
k=
k
A bound on the error in this approximation is
16 n^2
To ensure that |ET | 6 10
− 12 , it suffices to have
16 n^2
− 12 , i.e. n > 250 000
suffice.
(b)
0
1 + x^3 dx ≈
100 ∑
k=
2 k − 1
A bound on the error in this approximation is
32 n^2
To ensure that |EM | 6 10
− 12 , it suffices to have
32 n^2
− 12 , i.e. n > 125 000
suffice.
(c)
1
0
1 + x 3 dx ≈
50 ∑
k=
2 k − 1
49 ∑
k=
k
A bound on the error in this approximation is
320 n 4
− 9 .