ON THE NONLOCAL INITIAL VALUE PROBLEM

FOR FIRST ORDER DIFFERENTIAL SYSTEMS

Octavia Nica and Radu Precup
Department of Mathematics, BabesBolyai University,
400084 Cluj, Romania
Abstract
The paper is devoted to existence of solutions to initial value problems for
nonlinear rst order systems with nonlocal conditions. The proof will rely on the
Perov, Schauder and Leray-Schauder xed point principles which are applied to
a nonlinear integral operator splitted into two parts, one of Fredholm type for
the subinterval containing the points involved by the nonlocal condition, and
an another one of Volterra type for the rest of the interval. The novelty in this
paper is that this approach is combined with the technique that uses convergent
to zero matrices and vector norms.
Mathematics Subject Classication: 34A34, 34A12, 45G10.
Keywords and phrases: Nonlinear dierential system, nonlocal initial condition,
xed point theorem, vector norm, matrix convergent to zero.
1 Introduction
In this paper we deal with the nonlocal initial value problem for the rst order
dierential system
_

_
x

(t) = f (t, x(t) , y(t))

y

(t) = g (t, x(t) , y(t)) (a.e. on [0, 1])

x(0) +
m

k=1
a
k
x(t
k
) = 0
y (0) +
m

k=1
a
k
y(t
k
) = 0.
(1.1)
Here f, g : [0, 1] R
2
R are Carath`eodory functions, t
k
are given points with
0 t
1
t
2
... t
m
< 1 and a
k
, a
k
are real numbers with 1 +
m

k=1
a
k
= 0 and
1 +
m

k=1
a
k
= 0.
1
Notice that the nonhomogeneous nonlocal initial conditions
_

_
x(0) +
m

k=1
a
k
x(t
k
) = x
0
y (0) +
m

k=1
a
k
y(t
k
) = y
0
can always be reduced to the homogeneous ones (with x
0
= y
0
= 0) by the change of
variables x
1
(t) := x(t)a x
0
and y
2
(t) := y(t)a y
0
, where a =
_
1 +
m

k=1
a
k
_
1
and
a =
_
1 +
m

k=1
a
k
_
1
. According to [1], Problem (1.1) is equivalent to the following
integral system in C [0, 1]
2
:
_

_
x(t) = a
m

k=1
a
k
_
t
k
0
f (s, x(s) , y(s)) ds +
_
t
0
f (s, x(s) , y(s)) ds
y(t) = a
m

k=1
a
k
_
t
k
0
g (s, x(s) , y(s)) ds +
_
t
0
g (s, x(s) , y(s)) ds.
This can be viewed as a xed point problem in C [0, 1]
2
for the completely continuous
operator T = (T
1
, T
2
), T : C [0, 1]
2
C [0, 1]
2
, where T
1
and T
2
are given by
T
1
(x, y)(t) = a
m

k=1
a
k
_
t
k
0
f (s, x(s) , y(s)) ds +
_
t
0
f (s, x(s) , y(s)) ds,
T
2
(x, y)(t) = a
m

k=1
a
k
_
t
k
0
g (s, x(s) , y(s)) ds +
_
t
0
g (s, x(s) , y(s)) ds.
Operators T
1
and T
2
appear as sums of two integral operators, one of Fredholm
type, whose values depend only on the restrictions of functions to [0, t
m
], and the
other one, a Volterra type operator depending on the restrictions to [t
m
, 1] , as this
was pointed out in [2]. Thus, T
1
can be rewritten as T
1
= T
F
1
+ T
V
1
, where
T
F
1
(x, y)(t) =
_

_
a
m

k=1
a
k
_
t
k
0
f (s, x(s) , y(s)) ds +
_
t
0
f (s, x(s) , y(s)) ds, if t < t
m
a
m

k=1
a
k
_
t
k
0
f (s, x(s) , y(s)) ds +
_
t
m
0
f (s, x(s) , y(s)) ds, if t t
m
and
T
V
1
(x, y)(t) =
_
0, if t < t
m
_
t
t
m
f (s, x(s) , y(s)) ds, if t t
m
.
Similarly, T
2
= T
F
2
+ T
V
2
, where
2
T
F
2
(x, y)(t) =
_

_
a
m

k=1
a
k
_
t
k
0
g (s, x(s) , y(s)) ds +
_
t
0
g (s, x(s) , y(s)) ds, if t < t
m
a
m

k=1
a
k
_
t
k
0
g (s, x(s) , y(s)) ds +
_
t
m
0
g (s, x(s) , y(s)) ds, if t t
m
and
T
V
2
(x, y)(t) =
_
0, if t < t
m
_
t
t
m
g (s, x(s) , y(s)) ds, if t t
m
.
.
This allows us to split the growth condition on the nonlinear terms f(t, x, y) and
g(t, x, y) into two parts, one for t [0, t
m
] and another one for t [t
m
, 1] , in a such
way that one reobtains the classical growth when t
m
= 0, that is for the local initial
condition x(0) = 0. In what follows, the notation |x|
C[a,b]
stands for the max-norm
on C [a, b]
|x|
C[a,b]
= max
t[a,b]
|x(t)| ,
while x
C[a,b]
denotes the Bielecki norm
x
C[a,b]
=

x(t) e
(ta)

C[a,b]
for some suitable > 0.
Nonlocal initial value problems were extensively discussed in the literature by
dierent methods (see for example [1], [2], [4], [5], [7], [9]). The results in the present
paper extend to systems those established for equations in [2]. The method could be
adapted to treat systems of evolution equations as this was done for equations in [3].
In the next section three dierent xed point principles will be used in order to
prove the existence of solutions for the semilinear problem, namely the xed point
theorems of Perov, Schauder and Leray-Schauder (see [9]). In all three cases a key
role will be played by the so called convergent to zero matrices. A square matrix M
with nonnegative elements is said to be convergent to zero if
M
k
0 as k .
It is known that the property of being convergent to zero is equivalent to each of
the following three conditions (for details see [9], [10]):
(a) I M is nonsingular and (I M)
1
= I +M +M
2
+... (where I stands for
the unit matrix of the same order as M);
(b) the eigenvalues of M are located inside the unit disc of the complex plane;
(c) I M is nonsingular and (I M)
1
has nonnegative elements.
The following lemma whose proof is immediate from characterization (b) of con-
vergent to zero matrices will be used in the sequel:
3
Lemma 1.1 If A is a square matrix that converges to zero and the elements of an
other square matrix B are small enough, then A + B also converges to zero.
We nish this introductory section by recalling (see [9], [11]) three fundamental
results which will be used in the next sections. Let X be a nonempty set. By a
vector-valued metric on X we mean a mapping d : X X R
n
+
such that
(i) d(u, v) 0 for all u, v X and if d(u, v) = 0 then u = v;
(ii) d(u, v) = d(v, u) for all u, v X;
(iii) d(u, v) d(u, w) + d(w, v) for all u, v, w X.
Here, if x, y R
n
, x = (x
1
, x
2
, ..., x
n
), y = (y
1
, y
2
, ..., y
n
), by x y we mean x
i
y
i
for i = 1, 2, ..., n. We call the pair (X, d) a generalized metric space. For such a space
convergence and completeness are similar to those in usual metric spaces.
An operator T : X X is said to be contractive (with respect to the vector-
valued metric d on X) if there exists a convergent to zero matrix M such that
d(T(u), T(v)) Md(u, v) for all u, v X.
Theorem 1.2 (Perov) Let (X, d) be a complete generalized metric space and T :
X X a contractive operator with Lipschitz matrix M. Then T has a unique xed
point u

and for each u

0
X we have
d(T
k
(u
0
), u

) M
k
(I M)
1
d(u
0
, T(u
0
)) for all k N.
Theorem 1.3 (Schauder) Let X be a Banach space, D X a nonempty closed
bounded convex set and T : D D a completely continuous operator (i.e., T is
continuous and T(D) is relatively compact). Then T has at least one xed point.
Theorem 1.4 (LeraySchauder) Let (X, || . ||) be a Banach space, R > 0 and
T : B
R
(0; X) X a completely continuous operator. If ||u|| < R for every solution
u of the equation u = T(u) and any (0, 1), then T has at least one xed point.
Throughout the paper we shall assume that the following conditions are satised:
(H1) 1 +
m

k=1
a
k
= 0 and 1 +
m

k=1
a
k
= 0.
(H2) f, g : [0, 1] R
2
R is such that f(., x, y), g(., x, y) are measurable for
each (x, y) R
2
and f(t, ., .), g(t, ., .) are continuous for every t [0, 1].
2 Nonlinearities with the Lipschitz property.
Application of Perovs xed point theorem
Here we show that the existence of solutions to problem (1.1) follows from Perovs
xed point theorem in case that f, g satisfy Lipschitz conditions in x and y :
|f(t, x, y) f(t, x, y)|
_
b
1
|x x| +

b
1
|y y| if t [0, t
m
]
c
1
|x x| +c
1
|y y| if t [t
m
, 1] ,
(2.1)
4
|g(t, x, y) g(t, x, y)|
_
B
1
|x x| +

B
1
|y y| if t [0, t
m
]
C
1
|x x| +

C
1
|y y| if t [t
m
, 1]
(2.2)
for all x, y, x, y R.
Theorem 2.1 If f, g satisfy the Lipschitz conditions (2.1), (2.2) and matrix
M
0
:=
_
b
1
t
m
A
1

b
1
t
m
A
1
B
1
t
m
A
2

B
1
t
m
A
2
_
(2.3)
converges to zero, then problem (1.1) has a unique solution.
Proof. We shall apply Perovs xed point theorem in C [0, 1]
2
endowed with the
vector norm . dened by
u = (x , y)
for u = (x, y) , where for z C [0, 1] , we let
z = max
_
|z|
C[0,t
m
]
, z
C[t
m
,1]
_
.
We have to prove that T is a generalized contraction, more exactly that
T(u) T(u) M

u u
for all u = (x, y), u = (x, y) C [0, 1]
2
and some matrix M

converging to zero. To
this end, let u = (x, y), u = (x, y) be any elements of C [0, 1]
2
. For t [0, t
m
] , if we
denote
A
1
:= 1 +|a|
m

k=1
|a
k
| ,
we have
|T
1
(x, y)(t) T
1
(x, y)(t)|
=

a
m

k=1
a
k
_
t
k
0
f (s, x(s) , y(s)) ds +
_
t
0
f (s, x(s) , y(s)) ds
+a
m

k=1
a
k
_
t
k
0
f (s, x(s) , y(s)) ds
_
t
0
f (s, x(s) , y(s)) ds

A
1
_
t
m
0
|f (s, x(s) , y(s)) f (s, x(s) , y(s))| ds
b
1
t
m
A
1
|x x|
C[0,t
m
]
+

b
1
t
m
A
1
|y y|
C[0,t
m
]
.
Taking the supremum, we obtain that
|T
1
(x, y) T
1
(x, y)|
C[0,t
m
]
b
1
t
m
A
1
|x x|
C[0,t
m
]
+

b
1
t
m
A
1
|y y|
C[0,t
m
]
(2.4)
5
For t [t
m
, 1] and any > 0, we have
|T
1
(x, y)(t) T
1
(x, y)(t)|
=

a
m

k=1
a
k
_
t
k
0
f (s, x(s) , y(s)) ds +
_
t
0
f (s, x(s) , y(s)) ds
+a
m

k=1
a
k
_
t
k
0
f (s, x(s) , y(s)) ds
_
t
0
f (s, x(s) , y(s)) ds

b
1
t
m
A
1
|x x|
C[0,t
m
]
+

b
1
t
m
A
1
|y y|
C[0,t
m
]
+
_
t
t
m
(c
1
|x(s)x x(s)| +c
1
|y(s) y(s)| ) ds.
The last integral can be further estimated as follows:
_
t
t
m
(c
1
|x(s)x x(s)| +c
1
|y(s) y(s)| ) ds
= c
1
_
t
t
m
|x(s) x(s)| e
(st
m
)
e
(st
m
)
ds
+c
1
_
t
t
m
|y(s) y(s)| e
(st
m
)
e
(st
m
)
ds

c
1

e
(tt
m
)
x x
C[t
m
,1]
+
c
1

e
(tt
m
)
y y
C[t
m
,1]
.
Thus
|T
1
(x, y)(t) T
1
(x, y)(t)|
b
1
t
m
A
1
|x x|
C[0,t
m
]
+

b
1
t
m
A
1
|y y|
C[0,t
m
]
+
c
1

e
(tt
m
)
x x
C[t
m
,1]
+
c
1

e
(tt
m
)
y y
C[t
m
,1]
Dividing by e
(tt
m
)
and taking the supremum when t [t
m
, 1], we obtain
T
1
(x, y) T
1
(x, y)
C[t
m
,1]
(2.5)
b
1
t
m
A
1
|x x|
C[0,t
m
]
+

b
1
t
m
A
1
|y y|
C[0,t
m
]

c
1

x x
C[t
m
,1]
+
c
1

y y
C[t
m
,1]
.
Now (2.4) and (2.5) imply that
T
1
(x, y) T
1
(x, y)
_
b
1
t
m
A
1
+
c
1

_
x x +
_

b
1
t
m
A
1
+
c
1

_
y y . (2.6)
Similarly
T
2
(x, y) T
2
(x, y)
_
B
1
t
m
A
2
+
C
1

_
x x +
_

B
1
t
m
A
2
+

C
1

_
y y ,
(2.7)
6
where
A
2
= 1 +|a|
m

k=1
|a
k
| .
Using the vector norm we can put both inequalities (2.6), (2.7) under the vector
inequality
T (u) T (u) M

u u ,
where
M

=
_
b
1
t
m
A
1
+
c
1

b
1
t
m
A
1
+
ec
1

B
1
t
m
A
2
+
C
1

B
1
t
m
A
2
+
e
C
1

_
. (2.8)
Clearly matrix M

can be represented as M

= M
0
+ M
1
, where
M
1
=
_
c
1

ec
1

C
1

e
C
1

_
.
Since M
0
is assumed to be convergent to zero, from Lemma 1.1 we have that M

also converges to zero for large enough > 0.The result follows now from Perovs
xed point theorem.
3 Nonlinearities with growth at most linear.
Application of Schauders xed point theorem
Here we show that the existence of solutions to problem (1.1) follows from Schauders
xed point theorem in case that f, g satisfy instead of the Lipschitz condition the
more relaxed condition of growth at most linear:
|f(t, x, y)|
_
b
1
|x| +

b
1
|y| + d
1
if t [0, t
m
]
c
1
|x| +c
1
|y| + d
2
if t [t
m
, 1] ,
(3.1)
|g(t, x, y)|
_
B
1
|x| +

B
1
|y| + D
1
if t [0, t
m
]
C
1
|x| +

C
1
|y| + D
2
if t [t
m
, 1] .
(3.2)
Theorem 3.1 If f, g satisfy (3.1), (3.2) and matrix (2.3) converges to zero, then
problem (1.1) has at least one solution.
Proof. In order to apply Schauders xed point theorem, we look for a nonempty,
bounded, closed and convex subset B of C [0, 1]
2
so that T(B) B. Let x, y be any
elements of C [0, 1] .
For t [0, t
m
] , we have
|T
1
(x, y)(t)| =

a
m

k=1
a
k
_
t
k
0
f (s, x(s) , y(s)) ds +
_
t
0
f (s, x(s) , y(s)) ds

A
1
_
t
m
0
|f (s, x(s) , y(s))| ds
b
1
t
m
A
1
|x|
C[0,t
m
]
+

b
1
t
m
A
1
|y|
C[0,t
m
]
+ d
1
t
m
A
1
.
7
Taking the supremum, we obtain that
|T
1
(x, y)|
C[0,t
m
]
b
1
t
m
A
1
|x|
C[0,t
m
]
+

b
1
t
m
A
1
|y|
C[0,t
m
]
. (3.3)
For t [t
m
, 1] and any > 0, we have
|T
1
(x, y)(t)| =

a
m

k=1
a
k
_
t
k
0
f (s, x(s) , y(s)) ds +
_
t
0
f (s, x(s) , y(s)) ds

b
1
t
m
A
1
|x|
C[0,t
m
]
+

b
1
t
m
A
1
|y|
C[0,t
m
]
+ d
1
t
m
A
1
+
_
t
t
m
(c
1
|x(s)| +c
1
|y(s)| + d
2
) ds
b
1
t
m
A
1
|x|
C[0,t
m
]
+

b
1
t
m
A
1
|y|
C[0,t
m
]
+ d
1
t
m
A
1
+ (1 t
m
)d
2
+c
1
_
t
t
m
|x(s)| e
(st
m
)
e
(st
m
)
ds
+c
1
_
t
t
m
|y(s)| e
(st
m
)
e
(st
m
)
ds
b
1
t
m
A
1
|x|
C[0,t
m
]
+

b
1
t
m
A
1
|y|
C[0,t
m
]
+ c
0
+
c
1

e
(tt
m
)
x
C[t
m
,1]
+
c
1

e
(tt
m
)
y
C[t
m
,1]
,
where c
0
= d
1
t
m
A
1
+(1 t
m
)d
2
. Dividing by e
(tt
m
)
and taking the supremum, it
follows that
T
1
(x, y)
C[t
m
,1]
b
1
t
m
A
1
|x|
C[0,t
m
]
+

b
1
t
m
A
1
|y|
C[0,t
m
]
(3.4)
+
c
1

x
C[t
m
,1]
+
c
1

y
C[t
m
,1]
+ c
0
.
Clearly (3.3), (3.4) give
T
1
(x, y)
_
b
1
t
m
A
1
+
c
1

_
x +
_

b
1
t
m
A
1
+
c
1

_
y +c
0
(3.5)
where c
0
= max {d
1
t
m
A
1
, c
0
} . Similarly
T
2
(x, y)
_
B
1
t
m
A
2
+
C
1

_
x +
_

B
1
t
m
A
2
+

C
1

_
y +

C
0
, (3.6)
with

C
0
= max {D
1
t
m
A
2
, C
0
} . Now (3.5), (3.6) can be put together as
_
T
1
(x, y)
T
2
(x, y)
_
M

_
x
y
_
+
_
c
0

C
0
_
,
8
where matrix M

is given by (2.8) and converges to zero for a large enough > 0.

Next we look for two positive numbers R
1
, R
2
such that if x R
1
, y R
2
, then
T
1
(x, y) R
1
, T
2
(x, y) R
2
. To this end it is sucient that
_
_
_
_
b
1
t
m
A
1
+
c
1

_
R
1
+
_

b
1
t
m
A
1
+
ec
1

_
R
2
+c
0
R
1
_
B
1
t
m
A
2
+
C
1

_
R
1
+
_

B
1
t
m
A
2
+
e
C
1

_
R
2
+

C
0
R
2
. (3.7)
or equivalently
M

_
R
1
R
2
_
+
_
c
0

C
0
_

_
R
1
R
2
_
,
whence
_
R
1
R
2
_
(I M

)
1
_
c
0

C
0
_
.
Notice that I M

is invertible and its inverse (I M

)
1
has nonnegative elements
since M

converges to zero.
Thus, if B =
_
(x, y) C[0, 1]
2
: x R
1
, y R
2
_
, then T(B) B and Schauders
xed point theorem can be applied.
4 More general nonlinearities. Application of the Leray-
Schauder principle
We now consider that nonlinearlities f, g satisfy more general growth conditions,
namely:
|f(t, u)|
_

1
(t, |u|
e
) if t [0, t
m
]
(t)
1
(|u|
e
), if t [t
m
, 1],
(4.1)
|g(t, u)|
_

2
(t, |u|
e
) if t [0, t
m
]
(t)
2
(|u|
e
) if t [t
m
, 1],
(4.2)
for all u = (x, y) R
2
, where by |u|
e
we mean the euclidean norm in R
2
. Here

1
,
2
are Carath`eodory functions on [0, t
m
] R
+
, nondecreasing in their second ar-
gument, L
1
[t
m
, 1] , while
1
,
2
: R
+
R
+
are nondecreasing and 1/
1
, 1/
2

L
1
loc
(R
+
).
Theorem 4.1 Assume that conditions (4.1), (4.2) hold. In addition assume that
there exists a positive number R
0
such that for = (
1
,
2
) (0, )
2
_
1

1
_
t
m
0

1
(t, ||
e
)dt
1
A
1
1

2
_
t
m
0

2
(t, ||
e
)dt
1
A
2
implies ||
e
R
0
(4.3)
and
_

R

1
() +
2
()
>
_
1
t
m
(s)ds, (4.4)
9
where R

=
_
_
A
1
_
t
m
0

1
(t, R
0
)dt
_
2
+
_
A
2
_
t
m
0

2
(t, R
0
)dt
_
2
_
1/2
. Then problem (1.1)
has at least one solution.
Proof. The result will follow from the Leray-Schauder xed point theorem once
we have proved the boundedness of the set of all solutions to equations u = T(u),
for [0, 1]. Let u = (x, y) be such a solution. Then, for t [0, t
m
], we have
|x(t)| = |T
1
(x, y)(t)| (4.5)
=

a
m

k=1
a
k
_
t
k
0
f (s, x(s) , y(s)) ds +
_
t
0
f (s, x(s) , y(s)) ds

_
1 +|a|
m

k=1
|a
k
|
_
_
t
m
0
|f (s, x(s) , y(s))| ds
= A
1
_
t
m
0
|f (s, u(s))| ds
A
1
_
t
m
0

1
(s, |u(s)|
e
)ds.
Similarly
|y(t)| A
2
_
t
m
0

2
(s, |u(s)|
e
)ds. (4.6)
Let
1
= |x|
C[0,t
m
]
,
2
= |y|
C[0,t
m
]
. Then from (4.5), (4.6), we deduce
_

1
A
1
_
t
m
0

1
(t, ||
e
)dt

1
A
1
_
t
m
0

1
(t, ||
e
)dt.
This by (4.3) guarantees
||
e
R
0
. (4.7)
Next we let t [t
m
, 1]. Then
|x(t)| = |T
1
(x, y)(t)|
=

a
m

k=1
a
k
_
t
k
0
f (s, x(s) , y(s)) ds +
_
t
0
f (s, x(s) , y(s)) ds

A
1
_
t
m
0

1
(t, R
0
)dt +
_
t
t
m
|f (s, x(s) , y(s))| ds
A
1
_
t
m
0

1
(t, R
0
)dt +
_
t
t
m
(s)
1
(|u(s)|
e
)ds
= :
1
(t)
10
and similarly
|y(t)| A
1
_
t
m
0

2
(t, R
0
)dt +
_
t
t
m
(s)
2
(|u(s)|
e
)ds
= :
2
(t)
Denote (t) :=
_

2
1
(t) +
2
2
(t)
_
1/2
. Then
_

1
(t) = (t)
1
(|u(t)|
e
) (t)
1
( (t))

2
(t) = (t)
2
(|u(t)|
e
) (t)
2
( (t))).
. (4.8)
Consequently

(t) =

1
(t)

1
(t) +
2
(t)

2
(t)
(t)
(t)

1
(t)
(t)

1
( (t)) + (t)

2
(t)
(t)

2
( (t))
(t) [
1
( (t)) +
2
( (t))] .
It follows that
_
t
t
m

(s)

1
( (s)) +
2
( (s))
ds
_
t
t
m
(s)ds.
Furthermore, also using (4.4) we obtain
_
(t)
(t
m
)
d

1
() +
2
())

_
t
t
m
(s)ds
_
1
t
m
(s)ds <
_

R

1
() +
2
()
. (4.9)
Note that (t
m
) = R

. Then from (4.9) it follows that there exists R

1
such that
(t) R
1
for all t [t
m
, 1]. Then |x(t)| R
1
and |y(t)| R
1
, for all t [t
m
, 1], whence
|x|
C[t
m
,1]
R
1
, |y|
C[t
m
,1]
R
1
. (4.10)
Let R = max{R
0
, R
1
}. From (4.7), (4.10) we have |x|
C[0,1]
R and |y|
C[0,1]
R as
we wished.
Remark 4.2 If
1
(t, ) =
0
(t)
0
() , then the rst inequality in (4.3) implies
that
0
() c +c

for all R
+
and some constants c and c

, i.e. the growth of

0
is at most linear. However,
1
may have a superlinear growth. Thus we may say
that under the assumptions of Theorem 4.1, the growth of f (t, u) in u is at most
linear for t [0, t
m
] and can be superlinear for t [t
m
, 1]. The same can be said
about g (t, u) .
11
In particular, when t
m
= 0, that is when problem (1.1) becomes the classical
local initial value problem
_
_
_
x

= f (t, x, y)
y

= g (t, x, y) (a.e. t [0, 1])

x(0) = y (0) = 0,
(4.11)
our assumptions reduce to the classical conditions (see [6], [8]) and Theorem 4.1
gives the following result:
Corollary 4.3 Assume that
|f(t, u)| (t)
1
(|u|
e
),
|g (t, u)| (t)
2
(|u|
e
)
for t [0, 1] and u R
2
, where L
1
[0, 1] , while
1
,
2
: R
+
R
+
are
nondecreasing and 1/
1
, 1/
2
L
1
loc
(R
+
). In addition assume that
_

0
d

1
() +
2
()
>
_
1
0
(s)ds.
Then problem (4.11) has at least one solution.
References
[1] A. Boucherif, Dierential equations with nonlocal boundary conditions, Nonlin-
ear Anal. 47 (2001), 2419-2430.
[2] A. Boucherif and R. Precup, On the nonlocal initial value problem for rst order
dierential equations, Fixed Point Theory 4 (2003), 205-212.
[3] A. Boucherif and R. Precup, Semilinear evolution equations with nonlocal initial
conditions, Dynam. Systems Appl. 16 (2007), 507-516.
[4] L. Byszewski, Theorems about the existence and uniqueness of solutions of a
semilinear evolution nonlocal Cauchy problem, J. Math. Anal. Appl. 162 (1991),
494-505.
[5] L. Byszewski and V. Lakshmikantham, Theorem about the existence and unique-
ness of a solution of a nonlocal abstract Cauchy problem in a Banach space,
Appl. Anal. 40 (1990), 11-19.
[6] M. Frigon, Application de la th eorie de la transversalite topologique a des prob-
lemes non linearies pour des equations dierentielles ordinaires, Dissertationes
Math. 296, PWN, Warsawa, 1990.
[7] S.K. Ntouyas and P.Ch. Tsamatos, Global existence for semilinear evolution
equations with nonlocal conditions, J. Math. Anal. Appl. 210 (1997), 679-687.
12
[8] D. ORegan and R. Precup, Theorems of Leray-Schauder Type and Applications,
Gordon and Breach, Amsterdam, 2001.
[9] R. Precup, Methods in Nonlinear Integral Equations, Kluwer, Dordrecht, 2002.
[10] R. Precup, The role of matrices that are convergent to zero in the study of
semilinear operator systems, Math. Comp. Modelling 49 (2009), 703-708.
[11] R. P. Agarwal, M. Meehan, D. ORegan, Fixed Point Theory and Applications,
Cambridge University Press, Cambridge, 2001.
13