(9 intermediate revisions by 3 users not shown)
Line 1: Line 1:
 
== Homework 8 Collaboration Area ==
 
== Homework 8 Collaboration Area ==
  
Question on problem 15 in Sec 6.6.
+
Question on problem 15 in Sec 6.6:
  
 
I tried to obtain the expression for
 
I tried to obtain the expression for
Line 21: Line 21:
 
s/(s + 1)  *  1/(s+1)
 
s/(s + 1)  *  1/(s+1)
  
you'll need to compute the convolution integral:
+
via convolution, you'll need to compute the convolution integral:
  
 
<math>\int_0^t \cos(\tau)\sin(t-\tau)\ d\tau.</math>
 
<math>\int_0^t \cos(\tau)\sin(t-\tau)\ d\tau.</math>
  
 
You'll have to use a formula for the sine of the
 
You'll have to use a formula for the sine of the
difference of two angles and be very careful.
+
product of cosine and sine.  Here it is:
 +
 
 +
(cos A)(sin B) = (1/2)sin(A+B) + (1/2)sin(B-A).
 +
 
 +
You'll need to be very careful.
 
Remember,  t  acts like a constant in the integrals.
 
Remember,  t  acts like a constant in the integrals.
  
Line 33: Line 37:
 
of the book, a trig identity might be at fault.)
 
of the book, a trig identity might be at fault.)
  
* Another way to solve this problem is to recognize that the given expression is the derivative of 1 / [(s+2)^2 + 1]].....therefore greatly simplifying the solution (no trig identities required).   
+
Another, and perhaps better, way to solve this problem is to recognize that the given expression is the derivative of
 +
 
 +
1 / [(s+2)^2 + 1]]
 +
 
 +
...therefore greatly simplifying the solution (no trig identities required).
 +
 
 +
(You'll need to use the formula
 +
 
 +
L[ t f(t) ] = -F'(s)
 +
 
 +
to get the inverse transform.)  
 +
 
 +
P.257 #8: I'm a bit confused on how to get this problem started... Based on other problems in this section, we can factor out a t such that the function inside the transform is  t  times f(t) where f(t)=t^(n-1)e^(kt), but I don't see how that helps us much.
 +
 
 +
Answer from Bell:  You'll need to use mathematical induction for this problem.  Start with n=1.  Then use the formula
 +
 
 +
L[ t f(t) ] = -F'(s)
 +
 
 +
to reduce the n=2 case to the n=1 case, etc.
 +
 
 +
(AJ) Also, there's a better table of Laplace transforms at the end of chapter 6 that has the transform for f(t) = t^(n-1)e^(kt)
 +
 
 +
Yes, but if you wanted to compute the transform the easy way instead of using the differentiation formula from the section, you would probably use
 +
 
 +
L[ f(t)e^{at} ] = F(s-a)
 +
 
 +
to get
 +
 
 +
L[ t^n e^{kt} ] = n!/ (s-k)^{n+1}
 +
 
 +
because
 +
 
 +
L[ t^n ] = n!/s^{n+1}.
 +
 
 +
(That's how they got that formula in the table in an earlier section.)
  
 
[[2010 MA 527 Bell|Back to the MA 527 start page]]  
 
[[2010 MA 527 Bell|Back to the MA 527 start page]]  

Latest revision as of 14:21, 19 October 2010

Homework 8 Collaboration Area

Question on problem 15 in Sec 6.6:

I tried to obtain the expression for

s/(s + 1) * 1/(s+1)

but am not getting the correct result in the Laplace table of

t sin t.

I am using the convolution of cos(tau)*sin(t-tau). There is no t term in sight. Is it okay to read off the table? Even if it is, shouldn't the result be the same?

Answer:

To find the inverse Laplace transform of

s/(s + 1) * 1/(s+1)

via convolution, you'll need to compute the convolution integral:

$ \int_0^t \cos(\tau)\sin(t-\tau)\ d\tau. $

You'll have to use a formula for the sine of the product of cosine and sine. Here it is:

(cos A)(sin B) = (1/2)sin(A+B) + (1/2)sin(B-A).

You'll need to be very careful. Remember, t acts like a constant in the integrals.

There is only one correct answer, so you should get it that way. (If it looks different than the back of the book, a trig identity might be at fault.)

Another, and perhaps better, way to solve this problem is to recognize that the given expression is the derivative of

1 / [(s+2)^2 + 1]]

...therefore greatly simplifying the solution (no trig identities required).

(You'll need to use the formula

L[ t f(t) ] = -F'(s)

to get the inverse transform.)

P.257 #8: I'm a bit confused on how to get this problem started... Based on other problems in this section, we can factor out a t such that the function inside the transform is t times f(t) where f(t)=t^(n-1)e^(kt), but I don't see how that helps us much.

Answer from Bell: You'll need to use mathematical induction for this problem. Start with n=1. Then use the formula

L[ t f(t) ] = -F'(s)

to reduce the n=2 case to the n=1 case, etc.

(AJ) Also, there's a better table of Laplace transforms at the end of chapter 6 that has the transform for f(t) = t^(n-1)e^(kt)

Yes, but if you wanted to compute the transform the easy way instead of using the differentiation formula from the section, you would probably use

L[ f(t)e^{at} ] = F(s-a)

to get

L[ t^n e^{kt} ] = n!/ (s-k)^{n+1}

because

L[ t^n ] = n!/s^{n+1}.

(That's how they got that formula in the table in an earlier section.)

Back to the MA 527 start page

To Rhea Course List

Alumni Liaison

Recent Math PhD now doing a post-doctorate at UC Riverside.

Kuei-Nuan Lin