Dr. Z.

s Calc5 Lecture 2 Handout: The Inverse Laplace Transform and Derivatives

By Doron Zeilberger

Theory: The Laplace Transform is a dictionary that goes from functions of t (usually time) to
functions of s. It is often necessary to be able to translate back! Taking the little table of Lecture
1, and reversing it we get:
1
(a) L1 { } = 1
s
1 tk1
(b) L1 { } = (k = 1, 2, 3, ...),
sk (k 1)!
1
(c) L1 { } = eat ,
sa
1 sin kt
(d) L1 { }= ,
s2 + k 2 k
s
(e) L1 { 2 } = cos kt ,
s + k2
1 sinh kt
(f ) L1 { }= ,
s2 k 2 k
s
(g) L1 { 2 } = cosh kt .
s k2

Note that except for (a), each of these formulas contains infinitely many facts, since they involve
parameters. For example thanks to (b) we know the inverse-Laplace-Transform of 1/s, 1/s2 , 1/s3
etc.

Problem 2.1: Find L1 { 3s + 5

s5 }

Solution: By linearity:

3 5 1 1
L1 { + } = 3L1 { } + 5L1 { } .
s s5 s s5

By (a) (or (b) with k = 1)

1
L1 { } = 1 ,
s
By (c)
1
L1 { } = e5t .
s5
Combining, we have

3 5 1 1
L1 { + } = 3L1 { } + 5L1 { } = 3 1 + 5e5t = 3 + 5e5t .
s s5 s s5

Ans. to 2.1: L1 { 3s + 5
s5 } = 3 + 5e5t .

1
Problem 2.2: Find L1 { 2s+1
s2 +9 }

Sol.: Remember, we can always break-up the numerator!

2s + 1 2s 1 s 1
L1 { 2
} = L1 { 2 } + L1 { 2 } = 2L1 { 2 } + L1 { 2 } .
s +9 s +9 s +9 s +9 s +9
By (e) and (d) with k = 3, this equals

1
2 cos 3t + sin 3t .
3

Ans. to 2.2: L1 { 2s+1

s2 +9 } = 2 cos 3t +
1
3 sin 3t.

When we get complicated rational functions of s, we need to do a partial fraction decomposition.

Problem 2.3: Evaluate

3s 4
L1 { }
s2 3s + 2
Sol.: First we must factorize the denominator (unless it is already factored)

3s 4 3s 4
=
s2 3s + 2 (s 1)(s 2)

Now are are looking for magic numbers, lets call them A and B, such that

3s 4 A B
= + .
(s 1)(s 2) s1 s2

Adding the fractions on the right (using a/b + c/d = (ad + bc)/bd):

3s 4 A(s 2) + B(s 1)
= .
(s 1)(s 2) (s 1)(s 2)

The denominators automatically match, but to make this come true, we need the numerators to
match:
3s 4 = A(s 2) + B(s 1) .

Now we plug-in convenient values. When s = 2 we get

3 2 4 = A(2 2) + B(2 1) = 0 + B = B .

So B = 2. When s = 1 we get:

3 1 4 = A(1 2) + B(1 1) .

So
1 = A

2
and so A = 1. Once we know what A and B are we go back and write:
3s 4 1 2
= + .
(s 1)(s 2) s1 s2

So far it was just algebra. Only now are we ready to apply L1 .

3s 4 1 2 1 1
L1 { } = L1 { } + L1 { } = L1 { } + 2L1 { } .
s2 3s + 2 s1 s2 s1 s2
Using (c) with a = 1 and a = 2, we get:

et + 2e2t .

Ans. to 2.3:
3s 4
L1 { } = et + 2e2t .
s2 3s + 2

More Theory: The beauty of L is that it turns differential equations into algebraic equations!

If y(t) is any function of time, and if Y (s) is its Laplace Transform:

L{y(t)} = Y (s) ,

then the Laplace Transform of the first derivative, y 0 (t), is almost Y (s) multiplied by s, and if
y(0) = 0 then it is exactly that. We have:

L{y 0 (t)} = sY (s) y(0)

(Prove this!, Hint: Integration by parts.) Applying this very same formula to y 00 (t) we have

L{y 00 (t)} = s(sY (s) y(0)) y 0 (0) = s2 Y (s) sy(0) y 0 (0) .

One more time:

L{y 000 (t)} = s3 Y (s) s2 y(0) sy 0 (0) y 00 (0) ,

and so on. In general:

L{y (n) (t)} = sn Y (s) sn1 y(0) sn2 y 0 (0) . . . y (n1) (0) .

Equipped with this, we can solve Initial Value Problems.

Problem 2.4: Use Laplace Transform to solve the following initial-value problem.

y 0 + 2y = et , y(0) = 3 .

Solution: Let Y (s) = L{y(t)}. Applying L:

L{y 0 + 2y} = L{et } .

3
So
L{y 0 } + 2L{y} = L{et } .

But L{y 0 } = sY (s) y(0) = sY (s) + 3. Also L{et } = 1

s1 . So we have the algebraic equation:
(Lets abbrevaiate, and write Y for Y (s))

1
sY + 3 + 2Y = .
s1

Solving for Y , we get,

1 4 3s
(s + 2)Y = 3= .
s1 s1
Dividing both sides by (s + 2), we have an explicit expression for Y :

4 3s
Y (s) = .
(s 1)(s + 2)

In order to find y(t) we need to compute

4 3s
L1 { } .
(s 1)(s + 2)

Doing partial fractions:

4 3s A B
Y (s) = = +
(s 1)(s + 2) s1 s+2
We get
4 3s A(s + 2) + B(s 1)
=
(s 1)(s + 2) (s 1)(s + 2)
So
4 3s = A(s + 2) + B(s 1) .
1
When s = 1, we get 1 = 3A so A = 3 and when s = 2 we get 10 = B(3) so B = 10
3 , so the
simplified expression for Y is:
1
3 10
3
Y (s) = + .
s1 s+2
Now we are ready to take L1
1
10 1 1 10 1
y(t) = L1 { 3
+ 3
} = ( )L1 { } + ( )L1 { }=
s1 s+2 3 s1 3 s+2

1 t 10 2t
e e .
3 3
10 2t
Ans. to 2.4: y(t) = 13 et 3 e .

Problem 2.5: Use Laplace Transform to solve the following initial-value problem.

y 00 3y 0 + 2y = e3t , y(0) = 0 , y 0 (0) = 0 .

4
Solution: As usual L{y(t)} = Y (s). Applying L:

L{y 00 3y 0 + 2y} = L{e3t } , y(0) = 0 , y 0 (0) = 0 .

1
L{y 00 } 3L{y 0 } + 2L{y} =
s3
Note that L{y 0 } = sY y(0) = sY , L{y 00 } = s2 Y sy(0) y 0 (0) = s2 Y , so

1
s2 Y 3sY + 2Y = .
s3

1
(s2 3s + 2)Y = .
s3
1
(s 1)(s 2)Y = .
s3
Solving for Y :
1
Y = .
(s 1)(s 2)(s 3)
Partial Fractions:
1 A B C
= + + .
(s 1)(s 2)(s 3) s1 s2 s3
1 = A(s 2)(s 3) + B(s 1)(s 3) + C(s 1)(s 2) .

Conveninet values: s = 1: 1 = A(1)(2), so A = 21 ; s = 2: 1 = B(1)(1), so B = 1; s = 3:

1 = C(2)(1), so C = 12 . So
1 1
1
Y = 2 + + 2 .
s1 s2 s3
Finally,
1 1
1 1 1 1 1 1
y(t) = L1 {Y } = L1 { 2
+ + 2 } = L1 { + }
s1 s2 s3 2s1 s2 2s3

1 1 1 1 1 1 1 1
= L { } L1 { } + L1 { } = et e2t + e3t .
2 s1 s2 2 s3 2 2
Ans. to 2.5: y(t) = 12 et e2t + 21 e3t .