The Existence and Uniqueness Theorem (of the solution a first

order linear equation initial value problem)

Does an initial value problem always a solution? How many solutions are
there? The following theorem states a precise condition under which exactly
one solution would always exist for a given initial value problem.

Theorem: If the functions p and g are continuous on the interval I: α < t < β
containing the point t = t0, then there exists a unique function y = φ(t) that
satisfies the differential equation

y′ + p(t) y = g(t)

for each t in I, and that also satisfies the initial condition

y(t0) = y0,

where y0 is an arbitrary prescribed initial value.

That is, the theorem guarantees that the given initial value problem will
always have (existence of) exactly one (uniqueness) solution, on any interval
containing t0 as long as both p(t) and g(t) are continuous on the same interval.
The largest of such intervals is called the interval of validity of the given
initial value problem. In other words, the interval of validity is the largest
interval such that (1) it contains t0, and (2) it does not contain any
discontinuity of p(t) nor g(t). Conversely, neither existence nor uniqueness
of a solution is guaranteed at a discontinuity of either p(t) or g(t).

Note that, unless t0 is actually a discontinuity of either p(t) or g(t), there

always exists a non-empty interval of validity. If, however, t0 is indeed a
discontinuity of either p(t) or g(t), then the interval of validity will be empty.
Clearly, in such a case the conditions that the interval must contain t0 and
that it must not contain a discontinuity of p(t) or g(t) will be contradicting.

© 2008, 2012 Zachary S Tseng A-1 - 24

If so, such an initial value problem is not guaranteed to have a unique
solution at all.

Example: Consider the initial value problem solved earlier

cos(t) y′ − sin(t) y = 3t cos(t), y(2π) = 0.

The standard form of the equation is

y′ − tan(t) y = 3t

with p(t) = − tan(t) and g(t) = 3t. While g(t) is always continuous, p(t)
has discontinuities at t = ±π/2, ±3π/2, ±5π/2, ±7π/2, … According to
the Existence and Uniqueness Theorem, therefore, a continuous and
differentiable solution of this initial value problem is guaranteed to
exist uniquely on any interval containing t0 = 2π but not containing
any of the discontinuities. The largest such intervals is (3π/2, 5π/2).
It is the interval of validity of this problem. Indeed, the actual
solution y(t) = 3t tan(t) + 3 − 3sec(t) is defined everywhere within
this interval, but not at either of its endpoints.

© 2008, 2012 Zachary S Tseng A-1 - 25

How to find the interval of validity

For an initial value problem of a first order linear equation, the interval of
validity, if exists, can be found using this following simple procedure.

Given: y′ + p(t) y = g(t), y(t0) = y0.

1. Draw the number line (which is the t-axis).

2. Find all the discontinuities of p(t), and the discontinuities of g(t). Mark
them off on the number line.

3. Locate on the number line the initial time t0. Look for the longest
interval that contains t0, but contains no discontinuities.

Step 1: Draw the t-axis.

Step 2: Mark off the discontinuities.

Step 3: Locate t0 and determine the interval of validity.

© 2008, 2012 Zachary S Tseng A-1 - 26

Example: Consider the initial value problems
2 3t
(a) (t − 81) y′ + 5e y = sin(t), y(1) = 10π
2 3t
(b) (t − 81) y′ + 5e y = sin(t), y(10π) = 1

The equation is first order linear, so the theorem applies. The standard form
of the equation is
5e 3t sin(t )
y′ + 2 y= 2
t − 81 t − 81

5e 3t sin(t )
with p(t ) = and g ( t ) = . Both have discontinuities at t = ± 9.
t 2 − 81 t 2 − 81
Hence, any interval such that a solution is guaranteed to exist uniquely must
contain the initial time t0 but not contain either of the points 9 and −9.

In (a), t0 = 1, so the interval contains 1 but not ± 9. The largest such

interval is (−9, 9).

In (b), t0 = 10π, so the interval contains 10π but neither of ± 9. The

largest such interval is (9, ∞).

Remember that the value of y0 does not matter at all, t0 alone determines the
interval.

Suppose the initial condition is y(−100) = 5 instead. Then the largest

interval on which the initial value problem’s solution is guaranteed to exist
uniquely will be (−∞, −9).

Lastly, suppose the initial condition is y(−9) = 88. Then we would not be
assured of a unique solution at all. Since t = −9 is both t0 and a discontinuity
of p(t) and g(t). The interval of validity would be, therefore, empty.

© 2008, 2012 Zachary S Tseng A-1 - 27

Depending on the problem, the interval of validity, if exists, could be as
large as the entire real line, or arbitrarily small in length. The following
example is an initial value problem that has a very short interval of validity
for its unique solution.

Example: Consider the initial value problems

(t2 − 10−2000000) y′ + t y = 0, y(0) = α.

With the standard form

t
y' − y = 0,
t 2 − 10 −2000000

the discontinuities (of p(t)) are t = ±10−1000000. The initial time is t0 = 0.

Therefore, the interval of validity for its solution is the interval (−10−1000000,
10−1000000), an interval of length 2×10−1000000 units!

However, the important thing is that somewhere on the t-axis a unique

solution to this initial value problem exists. Different initial value α will
give different particular solution. But the solution will each uniquely exist,
at a minimum, on the interval (−10−1000000, 10−1000000).

Again, according to the theorem, the only time that a unique solution is not
guaranteed to exist anywhere is whenever the initial time t0 just happens to
be a discontinuity of either p(t) or g(t).

Now suppose the initial condition is y(0) = 0. It should be fairly easy to see
that the constant zero function y(t) = 0 is a solution of the initial value
problem. It is of course the unique solution of this initial value problem.
Notice that this solution exists for all values of t, not just inside the interval
(−10−1000000, 10−1000000). It exists even at discontinuities of p(t). This
illustrates that, while outside of the interval of validity there is no guarantee
that a solution would exist or be unique, the theorem nevertheless does not
prevent a solution to exist, even uniquely, where the condition required by
the theorem is not met.

© 2008, 2012 Zachary S Tseng A-1 - 28

Nonlinear Equations: Existence and Uniqueness of Solutions

A theorem analogous to the previous exists for general first order ODEs.

Theorem: Let the function f and ∂f ∂y be continuous in some rectangle α <

t < β, γ < y < δ, containing the point (t0, y0). Then, in some interval t0 − h < t
< t0 + h contained in α < t < β, there is a unique solution y = φ(t) of the initial
value problem
y′ = f (t,y), y(t0) = y0.

This is a more general theorem than the previous that applies to all first
order ODEs. It is also less precise. It does not specify a precise region that
a given initial value problem would have a solution or that a solution, when
it exists, is unique. Rather, it states a region that somewhere within there has
to be part of it in which a unique solution of the initial value problem will
exist. (It does not preclude that a second solution exists outside of it.)

The bottom line is that a nonlinear equation might have multiple solutions
corresponding to the same initial condition. On the other hand it is also
possible that it might not have a solution defined on parts of the region
where f and ∂f ∂y are both continuous.

Example: Consider the (nonlinear) initial value problem

y′ = t2 y 1/2, y(0) = 0.

When t = 0, ∂f ∂y is not continuous. Therefore, it would not necessarily

t6
have a unique solution. Indeed, both y = and y = 0 are functions that
36
satisfy the problem. (Verify this fact!)

© 2008, 2012 Zachary S Tseng A-1 - 29

Exercises A-1.2:

1 – 4 Find the general solution of each equation below.

1. y′ − t2 y = 4t2

2. y′ + 10 y = t2
1 t3
3. y ' − e y=0
t2
4. y′ − y = 2e t

5 – 15 Solve each initial value problem. What is the largest interval in

which a unique solution is guaranteed to exist?
5. y′ + 2y = t e−t, y(0) = 2

6. y′ − 11y = 4e 6t, y(0) = 9

7. t y′ − y = t2 + t, y(1) = 5

8. (t2 + 1) y′ − 2t y = t3 + t, y(0) = −4

9. y′ + (2t – 6t2) y = 0, y(0) = −8

2
10. t 2 y ′ + 4ty = , y(−2) = 0
t

11. (t2 − 49) y′ + 4t y = 4t, y(0) = 1 / 7

12. y′ − y = t2 + t, y(0) = 3

13. y′ + y = e t, y(0) = 1

14. t y′ + 4y = 4, y(−2) = 6

15. tan(t) y′ − sec(t) tan2(t) y = 0, y(0) = π

© 2008, 2012 Zachary S Tseng A-1 - 30

16 – 19 Without solving the initial value problem, what is the largest interval
in which a unique solution is guaranteed to exist for each initial condition?
(a) y(π) = 7, (b) y(1) = −9, (c) y(−4) = e.

(t − 8)(t − 1) t
16. (t + 5) y′ + y=
t −3 (t − 6)(t + 1)
t−2
17. t2 y′ + y = sec(t / 3)
t +3

18. (t2 + 4t − 5) y′ + tan(2t) y = t2 − 16

19. (4 − t2) y′ + ln(6 − t) y = e−t

20. Find the general solution of t2 y′ + 2t y = 2. Then show that both the
initial conditions y(1) = 1 and y(−1) = −3 result in an identical particular
solution. Does this fact violate the Existence and Uniqueness Theorem?

© 2008, 2012 Zachary S Tseng A-1 - 31

Answers A-1.2:
3
1. y = −4 + Ce t / 3
t2 t 1
2. y = − + + Ce −10 t
10 50 500
1 t  3
3. y = C exp e 
3 
4. y = 2t e + Ce t
t

5. y = t e−t − e−t + 3 e−2t, (−∞, ∞)

49 11t 4 6t
6. y = e − e , (−∞, ∞)
5 5
7. y = t2 + t ln t + 4t, (0, ∞)
8. y = (t 2 + 1)(ln( t 2 + 1) − 4) , (−∞, ∞)
9. y = −8 exp(2t3 – t2), (−∞, ∞)
1 4
10. y = 2 − 4 , (−∞, 0)
t t
t − 98t 2 + 343
4

11. y = , (−7, 7)
(t 2 − 49) 2
12. y = 6e t − t2 – 3t – 3, (−∞, ∞)
1 t 1 −t
13. y = e + e = cosh(t ) , (−∞, ∞)
2 2
14. y = 1 + 80t −4, (−∞, 0).
π sec( t )
15. y = e = πe sec( t )−1 , (−π/2, π/2)
e
16. (a) (3, 6); (b) (−1, 3); (c) (−5, −1).
17. (a) (0, 3π/2); (b) (0, 3π/2); (c) (−3π/2, −3).
18. (a) (3π/4, 5π/4); (b) no such interval exists; (c) (−5, −5π/4).
19. (a) (2, 6); (b) (−2, 2); (c) (−∞, −2).
2t + C 2t − 1
20. y = 2 ; they both have y = 2 as the solution; no, different initial
t t
conditions could nevertheless give the same unique solution.

© 2008, 2012 Zachary S Tseng A-1 - 32