1

3. LAPLACE TRANSFORMS

The Laplace Transform may be regarded as a mathematical device which is used to solve systems of ordinary
differential equations with constant coefficients. Such equations describe dynamical systems such as (i) a
group of masses coupled by means of springs and dampers, or (ii) a circuit with resistors and capacitances.
The manner in which such circuits may be transformed into a system of Ordinary Differential Equations or
directly into its Laplace Transform equivalent is beyond the scope of this unit, and is covered elsewhere. The
Laplace Transform is used very widely in Control Theory, but may also be used in heat transfer and fluid
mechanics. However, our aim here is merely to acquire some facility in using the transform itself and to use
it to solve some ODEs.

The taking of a Laplace Transform changes a function of time, y(t), say, into Y (s), a function of s, the
Laplace Transform variable. If one has a constant-coefficient ODE or system of such ODEs, the first step
is to perform a Laplace Transform of that equation. This process results in an algebraic system for the
transform of the dependent variable(s), which is then solved. Finally the process of taking the Inverse
Laplace Transform takes place; this results in a function of time which is the solution of the original system.
In other words, an ODE for f (t) is transformed into an algebraic equation for F (s) which is solved, and then
an inverse transform magically yields the required solution for f (t). For the engineer there is a ‘toolchest’ of
standard results which may be used to aid in the inversion process, and many of these results will be derived
in this section.

The Laplace Transform of the function f (t) is defined in the following way,
Z ∞
L[f (t)] = f (t)e−st dt = F (s). (3.1)
0

Note that the symbol L[ ] merely means ‘Laplace Transform of’ whatever happens to be in the square
bracket. Thus the Laplace Transform process changes the function f (t) into a new function F (s).

A natural question to ask is, is it possible to find f (t) if one is given a function F (s). In general, for the
applications that Engineers are interested in, the answer is ‘yes’. This will be done by means of look-up tables
of standard Laplace Transforms, and various common results/short-cuts. For the mathematician, however,
he/she could also go via the route of performing a suitable integral in the complex plane, the details of which
need not bother us. However, if we have a function F (s), then we write,

f (t) = L−1 [F (s)]

i.e. f (t) is the inverse Laplace Transform of F (s). The symbol L−1 [ ] means the ‘Inverse Laplace Transform
of’ whatever happens to be in the square bracket.

3.1. Some examples of Laplace Transforms.

Example 3.1 Find L[1].

This may be done simple by writing down the definition of the Laplace Transform and performing the
integration: Z ∞ h e−st i∞ 1
L[1] = 1 e−st dt = − = .
0 s 0 s
In this integration it is important to note that the answer is correct when s > 0. If s had been negative,
then the integral is infinite, Thus s > 0 is an existence condition for the transform. All Laplace Transforms
must have a range of values of s for which the integral exists. That being said, it is not often that one needs
to consider this aspect of Laplace Transforms.
 Mathematics 2 2

Example 3.2 Find L[e−at ].

Again, by definition, we have

∞ ∞
1
Z Z
L[e−at ] = e−at e−st dt = . e−(s+a)t dt =
0 0 s + a
This Laplace Transform exists when s + a > 0, i.e. when s > −a. When a = 0 we recover the result of
Example 3.1. When a = 1 and when a = −2 we get
1 1
L[e−t ] = and L[e2t ] = .
s+1 s−2
Hopefully it is easy to see that L[be−at ] = b/(s + a) so that multiplication of a function by a constant is
equivalent to multiplication of its Laplace Transform by the same constant.

Example 3.3 Find L[cos at].

By definition we have,
∞
s
Z
L[cos at] = cos at e−st dt =
2 + a2
.
0 s
The integral could have been performed using either integration by parts in the usual way, or by changing
cos at into eatj and taking the real part of the final answer. In addition,
a
L[sin at] = 2 .
s + a2

Example 3.4 Find L[t].

∞
1
Z
L[t] = t e−st dt = ,
0 s2
using one integration by parts.

Example 3.5 Find L[te−at ].

∞ ∞
1
Z Z
L[te−at ] = t e−at e−st dt = t e−(s+a)t dt = .
0 0 (s + a)2

3.2. Laplace Transforms of derivatives and the solution of some ODEs.

We haven’t yet derived a sufficient number of results for our eventual purposes, however we are almost in
a position where it is possible to solve some straightforward Ordinary Differential Equations. Therefore we
need now find the Laplace Transforms of some derivatives in order to enable us to solve such ODEs.

Example 3.6 Find L[y ′ (t)].

We start with the statement that L[y(t)] = Y (s). Using the definition of the LT we have,
Z ∞
′
L[y ] = y ′ e−st dt
0
h ih i∞ Z ∞ h i h i
−st
= y e − y −se−st dt
0
Z ∞ 0
= −y(0) + s y e−st dt [where ye−st → 0 as t → ∞]
0
= sY − y(0),
on using integration by parts once. So if the transform of y is Y , then the transform of y ′ is sY with
the initial condition for y(t) at t = 0 subtracted from it. Thus one might say that multiplication by s is
equivalent to a time derivative, but the initial condition mustn’t be forgotten.
 Mathematics 2 3

Example 3.7 Find L[y ′′ ].

Z ∞
L[y ′′ ] = y ′′ e−st dt
0
h ih i∞ h i h i∞ Z ∞ h ih i
′
= y e −st
y −se −st
+ y s2 e−st dt
0 0 0
= −y ′ (0) − sy(0) + s2 Y (s),

on using two integrations by parts. Here have also assumed that y ′ e−st tends to zero as t becomes infinite.

The equivalent expressions for the third and fourth derivatives may be shown to be

L[y ′′′ ] = −y ′′ (0) − sy ′ (0) − s2 y(0) + s3 Y (s)

and
L[y ′′′′ ] = −y ′′′ (0) − sy ′′ (0) − s2 y ′ (0) − s3 y(0) + s4 Y (s).

So we see that, apart from the values of y and and a sufficient number of derivatives of y evaluated at t = 0,
the Laplace Transform of an nth derivative is equivalent to multiplying Y by sn .

In addition, given that integration is the inverse process to differentiation, it might not come as a surprise
that,
hZ t i Y (s)
L y(τ ) dτ = .
0 s
In this result the variable τ is a dummy variable of integration. We shall not prove this result, but it may
be done using one integration by parts bearing in mind that,

d h t
Z i
y(τ ) dτ = y(t)
dt 0

precisely.

Example 3.8 Solve the ODE, y ′ + 2y = e−t , subject to the initial condition, y(0) = 0.

We can solve this equation by applying Laplace Transforms to each term in turn. We already know the
following:

L[y ′ ] = −y(0) + sY (s) Example 3.6

= sY as y(0) = 0

L[2y] = 2Y

1
L[e−t ] = . Example 3.2
s+1
Therefore the equation, together with its boundary condition, transforms immediately into the following
algebraic equation,
1
sY + 2Y = .
s+1
This may be solved for Y :
1
Y = ,
(s + 1)(s + 2)
which, after using partial fractions, becomes,
1 1
Y = − .
s+1 s+2
 Mathematics 2 4

Given the results in Example 3.2, we know that the function whose Laplace Transform corresponds to Y is

y = e−t − e−2t .

One may now check that this solution satisfies both the equation and its initial condition.

Example 3.9 Solve the equation, y ′ + y = e−t , subject to the initial condition, y(0) = c, where c is a
known, but unspecified, constant.

This equation is almost identical to that of Example 3.8, but the right hand side forcing term is proportional
to the Complementary Function of the ODE (i.e. repeated λ-values), and therefore the nature of the solution
will be different from that of Example 3.8. So it will be of interest to see how Laplace Transforms cope with
this special case. On taking Laplace Transforms of each term in the ODE we obtain,

1
sY − c + Y = ,
s+1

and therefore
1 c
Y = + .
(s + 1)2 s+1
The first term may be inverted using Example 3.5 with a = 1, and hence

y = te−t + ce−t .

In this solution the first term is the Particular Integral, while the second is the Complementary Function.

Example 3.10 Solve the equation y ′′ + 4y = 5e−t subject to y(0) = 0 and y ′ (0) = −1.

On using the formula for the Laplace Transform of a second derivative given in Example 3.7, we have

5
s2 Y − y ′ (0) − sy(0) + 4Y = ,
s+1

which may be rearranged into the form,

5
(s2 + 4)Y + 1 = ,
s+1

and therefore Y is given by,

5 1
Y = − .
(s + 1)(s2 + 4) s2 + 4
One may now proceed either by expanding the first fraction using the method of partial fractions, or else
combining the two fractions together to get (4 − s)/[(s + 1)(s2 + 4)], and then using the method of partial
fractions. Thus,
1 s
Y = − .
s + 1 s2 + 4
Results already derived (Ex. 3.2 with a = 1 and Ex. 3.3 with a = 2) are now sufficient to invert this
expression; we get
y = e−t − cos 2t.

3.3. Transfer Function.

Given the form of the last few examples we can generalize by making the right hand side forcing terms
arbitrary. Thus Examples 3.8 and 3.10 become,

y ′ + 2y = f (t) and y ′′ + 4y = f (t).

 Mathematics 2 5

On assuming that all initial conditions are zero, the Laplace Transforms of these equations yield,
 1   1 
Y = × F (s), and Y = 2 × F (s). (3.2)
s+2 s +4
In each of these cases the term in the round brackets is known as the Transfer Function. This function
characterizes the system which is modelled by the left hand side of the differential equation. It is of particular
importance to note that the inverse Laplace Transform of the transfer function is the ODE system’s response
to what is called the unit impulse, a function which will be described next.

3.4. The Unit Impulse.

Consider the function shown in Figure 3.1. The duration of this pulse is ǫ but its strength over that time
is ǫ−1 , and hence the total strength, which is the area under the graph, is precisely equal to 1. Hence this
may be called a unit pulse. The Laplace Transform of the unit pulse is
Z ∞
L[f (t)] = f (t) e−st dt
Z0 ǫ
= ǫ−1 e−st dt
0
1 h 1 − e−ǫs i
= . (3.3)
ǫ s

The unit impulse is what is obtained when ǫ becomes infinitesimally small. In the limit as ǫ → 0 the unit
pulse now has an infinite strength over an interval of length zero, but the total area remains equal to 1;
this is the unit impulse, which is also known as the delta function or, in physics contexts, as the Dirac delta
function. It is denoted by δ(t) and may be defined formally as,
δ(t) = lim f (t),
ǫ→0
where f (t) is as shown in Figure 3.1.

f (t)

1/ǫ

t
0 ǫ

Figure 3.1. Sketch of a unit pulse of duration ǫ and magnitude

1/ǫ. The total area is equal to 1.

The unit impulse has an important role in many areas of physics, mathematics and different branches of
engineering, and therefore it is of interest to determine its Laplace Transform. This may be done by letting
ǫ → 0 in (3.3) above. To do this we need to use the Taylor’s series expansion of e−ǫs :
(ǫs)2 (ǫs)3 (ǫs)4
e−ǫs = 1 − ǫs + − + ···. (3.4)
2! 3! 4!
 Mathematics 2 6

When (3.4) is substituted into (3.3) we get,

1 h 1 − (1 − ǫs + (ǫs)2 /2! − (ǫs)3 /3! · · ·) i

L[f (t)] =
ǫ s
ǫs − (ǫs)2 /2! + (ǫs)3 /3! · · ·
=
ǫs
ǫs[1 − ǫs/2! + (ǫs)2 /3! · · ·]
= ,
ǫs
= 1 − ǫs/2! + (ǫs)2 /3! · · · ,
→1 as ǫ → 0.

Therefore,
L[δ(t)] = 1.

This value could also have been obtained using L’Hôpital’s rule:

1 − e−ǫs se−ǫs
L[δ(t)] = lim = lim = 1,
ǫ→0 ǫs ǫ→0 s

on taking derivatives of both the numerator and denominator with respect to ǫ.

The unit impulse also has the property that,

Z ∞ Z ∞
g(t) δ(t) dt = g(0) δ(t) dt
−∞ −∞
since g(t) = g(0) where δ(t) is nonzero
Z ∞
= g(0) δ(t) dt elementary property of integrals
−∞
= g(0) × 1 = g(0).

In other words the integral of an function multiplied by the unit impulse is equivalent to picking out the
value of g(t) when t = 0. The apparent sleight of hand with the first equals sign above is motivated by the
fact that both g(t)δ(t − a) and g(a)δ(t − a) are equal to zero when t 6= a.

More generally, the unit impulse does not have to be located at x = 0. When it is centred at t = a it is
denoted by δ(t − a) and it is therefore infinite when (t − a) = 0. We now have,
Z ∞ Z ∞ Z ∞
g(t) δ(t − a) dt = g(a) δ(t − a) dt = g(a) δ(t − a) dt = g(a).
−∞ −∞ −∞

Following from this, the Laplace Transform of δ(t − a) is,

Z ∞
L[δ(t − a)] = δ(t − a) e−st dt = e−as .
0

Example 3.11 Solve y ′ + ay = δ(t) subject to y(0) = 0.

Again using known results we obtain the transformed version of the equation:

(s + a)Y = 1,

for which
1
Y = ⇒ y = e−at . (3.5)
s+a
 Mathematics 2 7

It is worth noting that while Y here is identical to the transfer function for the system, the y(t) that we
have found is known as the impulse response of the system.

It is also important to note that the solution for y(t) appears to violate the initial condition y(0) = 0, since
it is clear from (3.5) that y(0) = 1. This is an effect of having a unit impulse, and it does not contradict the
initial condition. Figure 3.2 compares three different solutions of y ′ + ay = f (t) with a = 1, where f (t) is
the unit pulse defined in Figure 3.1, but where ǫ = 0.5, 0.1 and 0. We see that, as ǫ → 0, y rises increasingly
sharply until, when ǫ = 0, it appears to be an instantaneous jump.

1.0

0.8

0.6

0.4

0.2

0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 t

Figure 3.2. Comparison between the solutions of y ′ +y = f (t) subject

to y(0) = 0 where f (t) is the unit pulse with ǫ = 0.5 (long dashes)
and ǫ = 0.1 (short dashes), and ǫ = 0 (i.e. unit impulse — continuous
curve.

Example 3.12 Solve the equation y ′′ + (a + b)y ′ + aby = δ(t) subject to y(0) = y ′ (0) = 0.

On taking the Laplace Transform of the equations we get,

 
s2 + (a + b)s + ab Y = 1
1
⇒ Y =
s2 + (a + b)s + ab
1 h 1 1 i
= − (partial fractions)
b−a s+a s+b
1 h −at i
⇒ y= e − e−bt
b−a
In this solution we may easily determine mathematically that y(0) = 0, but that y ′ (0) = 1. Thus the impulse
response for a second order ODE with zero initial conditions has a nonzero first derivative at t = 0. More
generally, for an nth order ODE with zero initial conditions, the unit impulse response appears to violate
the initial condition for the (n − 1)st derivative.

3.5. The Unit Step Function.

This is sketched in Figure 3.3a and is denoted either by u(t) or by H(t). The latter notation comes from its
alternative name, the Heaviside step function. Mathematically it is defined as
H(t) = 0 (t < 0) and H(t) = 1 (t > 0).
 Mathematics 2 8

An alternative representation is Z t
H(t) = δ(τ ) dτ.
−∞
Indeed, we may also write
dH(t)
δ(t) = .
dt
The step rise does not need to occur at t = 0. Its equivalent at t = a is denoted by H(t − a) and this is
shown in Figure 3.3b for the case a = 1.

4
H(t)

-1
-2 -1 0 1 2 t

Figure 3.3a. Depicting the unit (or Heaviside) step function H(t).

4
H(t − 1)

-1
-2 -1 0 1 2 t

Figure 3.3b. Depicting the unit (or Heaviside) step function H(t−1).

The Laplace Transform of H(t − a) is given by

Z ∞
L[H(t − a)] = H(t − a) e−st dt
0
 Mathematics 2 9
Z a Z ∞
= 0 × e−st dt + 1 × e−st dt splitting the integral
0 a
Z ∞
=0+ 1 × e−st dt
a
h e−st i∞ e−as
= − = .
s a s
We will need both the unit impulse and the unit step function later.

3.6. The shift theorem in s.

Given that Z ∞
L[f (t)] = f (t) e−st dt = F (s),
0

then
Z ∞
L[f (t)e−at ] = f (t)e−at e−st dt
Z0 ∞
= f (t)e−(s+a)t dt
0
= F (s + a).

This is the shift theorem in s. It say, in effect, that the multiplication of a function of time changes the
apparent origin of its Transform.

Example 3.13 Find the Laplace Transform of te−at .

Given that L[t] = 1/s2 (see Example 3.4), the shift theorem tells us immediately that

L[te−at ] = 1/(s + a)2 .

Example 3.14 Find the Laplace Transform of cos bt e−at .

Example 3.3 shows that L[cos bt] = s/(s2 + b2 ), and hence the shift theorem gives,

s+a
L[cos bt e−at ] = .
(s + a)2 + b2

We may also use this shift theorem to find inverse transforms. Indeed, this is its main use. The following is
a typical example.
s
Example 3.15 Find the inverse transform of .
s2 + 4s + 13
Given the form on the denominator in Example 3.14, we need to write s2 + 4s + 13 in a similar form by
completing the square. Thus,

s2 + 4s + 13 = (s + 2)2 + 9 = (s + 2)2 + 32 .

Therefore we can decompose the given function of s as follows:

s s
= on completing square
s2 + 4s + 13 (s + 2)2 + 32
s+2 2
= 2 2
− to get (s + 2) everywhere
(s + 2) + 3 (s + 2)2 + 32
s+2 2h 3 i
= − .
(s + 2)2 + 32 3 (s + 2)2 + 32
 Mathematics 2 10

This last step involved placing a 3 in the numerator of the fraction. The reason behind the various moves
above is that we are trying to force our original function into a form which is as close as possible to the
Laplace Transform results that
s 3
L[cos 3t] = and L[sin 3t] = .
s2 + 3 2 s2 + 3 2
Therefore the s-shift theorem tells us that
h s i h s+2 i
2 −1
h 3 i
L−1 2 = L−1 − 3 L
s + 4s + 13 (s + 2)2 + 32 (s + 2)2 + 32
= e−2t cos 3t − 32 e−2t sin 3t.

3.7. The shift theorem in t.

Consider first the difference between the functions f (t) and H(t − a)f (t − a) as demonstrated by Figure 3.4,
which shows a unit shift (a = 1). From the point of view of the Laplace Transform, we are not interested
in how f (t) varies for negative values of t because the Laplace Transform integral takes place over positive
values of t only. In Figure 3.4 we see f (t) shifted in time by an amount a, but where that part of f (t)
corresponding to t < 0 has been replaced by zero; this is why we have used the unit step function. Our aim
here is to relate the Laplace Transform of H(t − a)f (t − a) to that of f (t).

1.00

0.75

0.50

0.25

0.00

-0.25

-0.50

-0.75

-1.00
-1 0 1 2 3 4 t

Figure 3.4. Depicting the functions f (t) (solid curve) and H(t−1)f (t−1) (dashed
curve). Also shown is the variation of f (t) for negative values of t (dotted curve),
but these values are of no relevance to Laplace Transforms because integration
is over positive values of t only.

By definition we have
Z ∞
L[H(t − a)f (t − a)] = f (t − a) H(t − a) e−st dt
0
Z ∞
= f (t − a) H(t − a) e−st dt as the integrand is zero in 0 ≤ t < a
a
Z ∞
= f (t − a) e−st dt as H(t − a) = 1 in a ≤ t < ∞
a
Z ∞
= f (τ )e−s(τ +a) dτ using τ = t − a
0
Z ∞
−sa
=e f (τ ) e−sτ dτ e−sa is a constant with respect to τ
0
= e−sa F (s).
 Mathematics 2 11

Therefore we have shown that

L[H(t − a)f (t − a)] = e−sa F (s),

which is the Shift Theorem in t.

e−as
Example 3.16 What is the inverse transform of ?
s2
We know that L[t] = 1/s2 , and therefore the Shift Theorem in t tells us that
h e−as i
L−1 = H(t − a) × (t − a),
s2
as shown below.

Example 3.17 Find the inverse Laplace Transform of e−as /(s + b)2 .

This is an example of the use of both shift theorems. The exponential tells us that we need the t-shift
theorem. The presence of the (s + b) tells us that the s-shift is needed. From all the mess we can see that
the underlying function is 1/s2 . We know that L[t] = 1/s2 , and therefore the Shift Theorem in s tells us
that
1
L[te−bt ] = .
(s + b)2
Now the Shift Theorem in t tells us that

e−as
L[H(t − a) × (t − a)e−b(t−a) ] = ,
(s + b)2

where we have meticulously changed every t in te−bt to (t − a). Thus

h e−as i
L−1 = H(t − a) × (t − a)e−b(t−a) .
(s + b)2

This function of time (with a = b = 1) is illustrated in Figure 3.4 as the dashed curve.

Note that it is also possible to derive this solution by applying the s-shift first and the t-shift second — I’ll
leave that as an exercise, but it is slightly more awkward to do.

3.8. Convolution Theorem.

Given the two functions f (t) and g(t) and their respective transforms, F (s) and G(s), then the convolution
of f and g is given by Z t Z t
f ∗g = f (τ )g(t − τ ) dτ = f (t − τ )g(τ ) dτ,
0 0

where it should be noted that each integral results in the same function of t. We may say that this function
is “the convolution of f and g”.

The main reason we consider such a bizarre looking function is that its Laplace Transform is not only simple
but useful. The result we need is
L[f ∗ g] = F (s) G(s).

Therefore the Laplace Transform of the convolution of two functions is simply the product of their respective
transforms. The proof of this is given below, but it will not be asked for in the end-of-semester examinations.

Proof: We start by writing down the definition of the Laplace Transform:

Z ∞ hZ t i
L[f ∗ g] = f (τ )g(t − τ ) dτ e−st dt.
0 0
 Mathematics 2 12

This nested pair of integrals may be simplified by first reversing the order of integration. As the inner
integral is a τ -integral, we see that, for a chosen value of t, τ runs from 0 to t. If we were to reverse the
order of integration, then for a chosen value of τ we must have t running from τ to ∞. Therefore the above
expression becomes, Z Z ∞ ∞
L[f ∗ g] = f (τ )g(t − τ )e−st dt dτ.
0 τ

Let us now change variable from t to t̂ where t̂ = t − τ . Thus dt̂ = dt, and the lower limit in the inner
integral will change from τ to zero (this is because t = τ ⇒ t̂ = 0). So we get
Z ∞Z ∞
L[f ∗ g] = f (τ )g(t̂)e−s(t̂+τ ) dt̂ dτ
0 0
Z ∞Z ∞
= f (τ )e−sτ g(t̂)e−st̂ dt̂ dτ
Z0 ∞ 0 Z ∞
= f (τ )e−sτ dτ g(t̂)e−st̂ dt̂
0 0
= F (s) G(s).

Example 3.18 Given that L[t] = 1/s2 use the convolution theorem to find L−1 [1/s4 ].

Given that L[f ∗g] = F (s)G(s) we clearly have L−1 [F (s)G(s)] = f ∗g. We may therefore choose F = G = 1/s2
and therefore f = g = t, from Example 3.4.

Our solution is, therefore,

Z t
t∗t= τ × (t − τ ) dτ,
0
Z t
= (τ t − τ 2 ) dτ,
0
h it
1 2
= 2τ t − 31 τ 3
0
= 16 t3 .

Therefore L[ 16 t3 ] = 1/s4 .

Example 3.19 Use the convolution theorem to find the inverse Laplace Transform of

1 1
× .
s+1 s+2

We already know that L[e−t ] = 1/(s + 1) and L[e−2t ] = 1/(s + 2). Therefore if we set f (t) = e−t and
g(t) = e−2t , then the function we want is
Z t
f ∗g =e −t
∗e −2t
= e−τ × e−2(t−τ ) dτ
0
Z t
= e−2t eτ dτ
0
Z t
=e −2t
eτ dτ
0
h it
=e eτ
−2t

h 0 i
−2t t
=e e −1 = e−t − e−2t .

Therefore we have obtained the final result of Example 3.8 using an alternative means.
 Mathematics 2 13

Finally, on referring back to the solutions given in Eqs. (3.2), the convolution theorem may now be applied,
and therefore we find that

y ′ + 2y = f (t) subject to y(0) = 0 has the solution y = e−2t ∗ f (t)

h i
1
y ′′ + 4y = f (t) subject to y(0) = y ′ (0) = 0 has the solution y= 2 sin 2t ∗ f (t),

because,
1 1
L[e−2t ] = and L[ 12 sin 2t] = .
s+2 s2 +4

Finally, finally, the solution given above to y ′ + 2y = f (t) subject to y(0) = 0 may be expanded in the
following way:

y = e−2t ∗ f (t) the solution above

Z t
= e−2(t−τ ) f (τ ) dτ using the convolution integral
0
Z t
=e −2t
e2τ f (τ ) dτ.
0

This final expression is what one would obtain if one were to solve y ′ + 2y = f (t) using the Integrating Factor
approach given in the ODEs section of this unit. I’ll leave that as an exercise....