Skip to main content

Advertisement

Solvability of a nonlocal boundary value problem for linear functional differential equations

Article metrics

  • 1948 Accesses

Abstract

In the paper, the problem on the existence and uniqueness of a solution to the nonlocal problem

u (t)=(u)(t)+q(t),u(a)=h(u)+c

is considered, where :C([a,b];R)L([a,b];R) and h:C([a,b];R)R are linear bounded operators, qL([a,b];R), and cR.

MSC:34K06, 34K10.

1 Introduction and notation

On the interval [a,b], we consider the boundary value problem

u (t)=(u)(t)+q(t),
(1)
u(a)=h(u)+c,
(2)

where :C([a,b];R)L([a,b];R) and h:C([a,b];R)R are linear bounded operators, qL([a,b];R), and cR. By a solution to the equation (1), we understand an absolutely continuous function u:[a,b]R satisfying equality (1) almost everywhere on the interval [a,b]. A solution to equation (1) satisfying the boundary condition (2) is said to be a solution to problem (1), (2).

The question on the solvability of various types of boundary value problems for functional differential equations and their systems is a classical topic in the theory of differential equations (see, e.g., [116] and references therein). There is a lot of interesting general results, but only a few efficient conditions are known, namely, in the case where a nonlocal boundary condition is considered. In the present paper, new efficient conditions are found sufficient for the unique solvability of problem (1), (2). An important particular case of the boundary condition (2) is

u(a)=λu(b)+c
(3)

with λR, which in turn contains the initial condition (if λ=0), the periodic condition (if λ=1), and the anti-periodic condition (if λ=1). Problem (1), (3) is studied, e.g., in [13, 1719]. In [20, 21], the first step of our investigation in the general case was done. It is very useful to consider the boundary condition (2) as a nonlocal perturbation of the two-point condition (3). Therefore, we assume throughout the paper that the functional h is defined by the formula

h(v) = def λv(b)+ h 0 (v) h 1 (v)for vC ( [ a , b ] ; R ) ,
(4)

where λ>0 and h 0 , h 1 P F a b . There is no loss of generality in assuming this, because an arbitrary functional h can be represented in form (4).

The paper is organized as follows. Main results are formulated and proved in Section 2. In Section 3, the main results are applied to the equation with argument deviations

u (t)=p(t)u ( τ ( t ) ) g(t)u ( μ ( t ) ) +q(t),
(5)

where p,gL([a,b]; R + ), qL([a,b];R), and τ,μ:[a,b][a,b] are measurable functions. Some sufficient conditions for the validity of the inclusion V ˜ a b (h), which are part of the conditions for the main results, are given in Section 4.

The following notation is used throughout the paper:

  1. 1.

    is the set of all real numbers, R + =[0,+[.

  2. 2.

    C([a,b];R) is the Banach space of continuous functions v:[a,b]R endowed with the norm v C =max{|v(t)|:t[a,b]}.

  3. 3.

    C ˜ ([a,b];D), where DR, is the set of absolutely continuous functions v:[a,b]D.

  4. 4.

    L([a,b];R) is the Banach space of Lebesgue integrable functions p:[a,b]R endowed with the norm p L = a b |p(s)|ds.

  5. 5.

    L([a,b];D)={pL([a,b];R):p:[a,b]D}, where DR.

  6. 6.

    C([a,b];D)={vC([a,b];R):v:[a,b]D}, where DR.

  7. 7.

    L a b is the set of linear bounded operators :C([a,b];R)L([a,b];R). P a b is the set of operators L a b , mapping the set C([a,b]; R + ) into the set L([a,b]; R + ).

  8. 8.

    F a b is the set of linear bounded functionals h:C([a,b];R)R. P F a b is the set of functionals h F a b mapping the set C([a,b]; R + ) into the set R + .

  9. 9.

    C h ([a,b];R)={vC([a,b];R):v(a)=h(v)}, where h F a b .

2 Main results

We assume throughout the paper that the following assumptions hold:

(H1) If h(1)=1, then the operator is supposed to be ‘nontrivial’ in the sense that the condition (1)0 holds.

(H2) h ˜ 0, where the functional h ˜ is defined by the formula h ˜ (v)=h(v)v(a) for vC([a,b];R).

Since we are interested in the unique solvability of problem (1), (2) for every q and c, both hypotheses (H1) and (H2) are rather natural. Indeed, if (1)0, then an arbitrary constant function is a solution to problem (1), (2) with q0 and c=0 in the case, where h(1)=1. On the other hand, the assumption (H2) guarantees that the boundary condition (2) is not ‘degenerated.’

Before formulation of the main results, we introduce the following definitions.

Definition 2.1 [22]

Let h F a b . An operator L a b is said to belong to the set V ˜ a b (h), if every function u C ˜ ([a,b];R), satisfying the relations

u (t)(u)(t)for a.e. t[a,b],u(a)h(u)

is nonpositive on the interval [a,b].

Definition 2.2 [23]

An operator L a b is said to belong to the set S a b (a) (resp. S a b (b)) if every function u C ˜ ([a,b];R) satisfying the relations

u (t)(u)(t)for a.e. t[a,b],u(a)0 ( resp.  u ( b ) 0 )

is nonnegative (resp. nonpositive) on the interval [a,b].

Remark 2.1 Efficient conditions, guaranteeing the validity of the inclusions V ˜ a b (h) and S a b (a), S a b (b), are stated, respectively, in [22] and [23].

2.1 Formulation of results

For the sake of transparency, we first formulate all the results; their proofs are postponed till Section 2.2 below.

Theorem 2.1 Assume that there exist operators φ 0 V ˜ a b (h) and φ 1 P a b such that the inequality

| ( v ) ( t ) φ 0 ( v ) ( t ) | φ 1 ( | v | ) (t)for a.e.t[a,b]
(6)

holds on the set C h ([a,b];R). If, moreover,

φ 0 φ 1 V ˜ a b (h),
(7)

then problem (1), (2) has a unique solution.

Corollary 2.1 Let = 0 1 with 0 , 1 P a b and the relation h(1)>1 hold. Moreover, there exists ε[0,1/2] such that

ε 0 V ˜ a b (h),(12ε) 0 1 V ˜ a b (h).
(8)

Then problem (1), (2) has a unique solution.

Remark 2.2 Choosing a suitable number ε in Corollary 2.1 and using the results established in [22], we can obtain several efficient conditions, sufficient for the unique solvability of problem (1), (2). However, we do not formulate them in detail. We note only that for ε= 1 2 , the assumption (8) has the form

1 2 0 V ˜ a b (h), 1 V ˜ a b (h).

Theorem 2.2 Let there exist φ V ˜ a b (ω) such that the inequality

(v)(t)sgnv(t)φ ( | v | ) (t)for a.e.t[a,b]
(9)

holds on the set C h ([a,b];R), where the functional ω is given by the formula

ω(v) = def λv(b) h 0 (v) h 1 (v)forvC ( [ a , b ] ; R ) .
(10)

Then problem (1), (2) has a unique solution.

Theorem 2.3 Let = 0 1 with 0 , 1 P a b and the relations

h(1)>1, h 0 (1)1
(11)

be fulfilled. Moreover, there exists a function γ C ˜ ([a,b];]0,+[) satisfying the conditions

γ (t) 1 (γ)(t) 0 (1)(t)for a.e.t[a,b],
(12)
γ(a)<h(γ),
(13)
γ(a)γ(b)< ω 1 ,
(14)

where

ω 1 =1+ ( 1 h 0 ( 1 ) ) min { 1 , 1 λ } +2 ( λ h 1 ( 1 ) ) min { 1 , 1 λ } .
(15)

Then problem (1), (2) has a unique solution.

Theorem 2.4 Let = 0 1 with 0 , 1 P a b and the relations

λ h 1 (1)>1, h 0 (1)>0
(16)

be fulfilled. Moreover, there exists a function γ C ˜ ([a,b];]0,+[) such that condition (12) is satisfied and

γ(a)λγ(b) h 1 (γ) h 0 (1),
(17)
γ(a)γ(b)< ω 2 ,
(18)

where

ω 2 =1+ 1 h 0 ( 1 ) λ +2 1 1 λ h 1 ( 1 ) .
(19)

Then problem (1), (2) has a unique solution.

Remark 2.3 The assumption h 0 (1)1 appearing in Theorem 2.3 is not supposed in Theorem 2.4. On the other hand, assumption (17) of Theorem 2.4 is stronger than assumption (13) of Theorem 2.3.

Theorem 2.5 Let P a b , the relations

h 0 (1)1, h 1 (1)<λ
(20)

hold, and there exists a function γC([a,b];R) satisfying the conditions

γ (t)(γ)(t)for a.e.t[a,b],
(21)
γ(a)<h(γ).
(22)

Let, moreover, at least one of the following conditions be fulfilled

  1. (a)
    a b (1)(s)ds< ω 1 ,
    (23)

where the number ω 1 is given by formula (15);

  1. (b)
    S a b (a);
    (24)
  2. (c)
    S a b (b).
    (25)

Then problem (1), (2) has a unique solution.

Remark 2.4 If the relation h(1)1 is fulfilled, then the assumption concerning the existence of a function γ in Theorem 2.5 can be omitted. Indeed, since the operator is supposed to be nontrivial in the case where h(1)=1, the function

γ(t)=1+ a t (1)(s)dsfor t[a,b]

satisfies conditions (21) and (22).

Remark 2.5 Define the operator φ:C([a,b];R)C([a,b];R) by setting

φ(w)(t) = def w(a+bt)for t[a,b],wC ( [ a , b ] ; R ) .

Let

ˆ ( w ) ( t ) = def ( φ ( w ) ) ( a + b t ) for a.e.  t [ a , b ]  and let all  w C ( [ a , b ] ; R ) , h ˆ ( w ) = def 1 λ w ( b ) 1 λ h 0 ( φ ( w ) ) + 1 λ h 1 ( φ ( w ) ) for  w C ( [ a , b ] ; R ) , q ˆ ( t ) = q ( a + b t ) for a.e.  t [ a , b ] , c ˆ = 1 λ c .

It is not difficult to verify that if u is a solution to problem (1), (2), then the function v = def φ(u) is a solution to the problem

v (t)= ˆ (v)(t)+ q ˆ (t),v(a)= h ˆ (v)+ c ˆ ,
(26)

and vice versa, if v is a solution to problem (26), then the function u = def φ(v) is a solution to problem (1), (2).

Using this transformation, we can immediately derive other conditions for the unique solvability of problem (1), (2), complementing those stated above. For example, Theorem 2.3 yields.

Theorem 2.3′ Let = 0 1 with 0 , 1 P a b and the relations

h(1)<1, h 1 (1)λ

be fulfilled. Let, moreover, there exist a function γ C ˜ ([a,b];]0,+[) satisfying the conditions

γ ( t ) 0 ( γ ) ( t ) + 1 ( 1 ) ( t ) for a.e. t [ a , b ] , γ ( a ) > h ( γ ) , γ ( b ) γ ( a ) < ω 3 ,

where

ω 3 =1+ ( λ h 1 ( 1 ) ) min { 1 , 1 λ } +2 ( 1 h 0 ( 1 ) ) min { 1 , 1 λ } .

Then problem (1), (2) has a unique solution.

2.2 Proofs

The following lemma is well known from the general theory of boundary value problems for functional differential equations (see, e.g., [15, 24]; in the case, where the operator is strongly bounded, see also [1, 3, 14]).

Lemma 2.1 Problem (1), (2) is uniquely solvable if and only if the corresponding homogeneous problem

u (t)=(u)(t),
(27)
u(a)=h(u),
(28)

has only the trivial solution.

Remark 2.6 It follows immediately from Definition 2.1 and Lemma 2.1 that under the condition V ˜ a b (h) problem (1), (2) has a unique solution for every qL([a,b];R) and cR.

Now, we are in position to prove the main results. According to Lemma 2.1, it is sufficient to show that the homogeneous problem (27), (28) has only the trivial solution.

Proof of Theorem 2.1 Let u be a solution to problem (27), (28). Then, in view of (6), we get

u ( t ) = φ 0 ( u ) ( t ) + ( u ) ( t ) φ 0 ( u ) ( t ) φ 0 ( u ) ( t ) + φ 1 ( | u | ) ( t ) for a.e.  t [ a , b ] ,
(29)
u ( t ) = φ 0 ( u ) ( t ) + ( u ) ( t ) φ 0 ( u ) ( t ) φ 0 ( u ) ( t ) φ 1 ( | u | ) ( t ) for a.e.  t [ a , b ] .
(30)

By virtue of the assumption φ 0 V ˜ a b (h) and Remark 2.6, the problem

α (t)= φ 0 (α)(t) φ 1 ( | u | ) (t),
(31)
α(a)=h(α)
(32)

has a unique solution α. It follows from relations (29)-(31) that

( u + α ) ( t ) φ 0 ( u + α ) ( t ) for a.e.  t [ a , b ] , ( u α ) ( t ) φ 0 ( u α ) ( t ) for a.e.  t [ a , b ] .
(33)

On the other hand, conditions (28) and (32) yield

(u+α)(a)=h(u+α),(uα)(a)=h(uα).
(34)

Therefore, by virtue of the assumption φ 0 V ˜ a b (h), relations (33) and (34) imply

| u ( t ) | α(t)for t[a,b].
(35)

Now, in view of (35) and the assumption φ 1 P a b , we get from (31) the relation

α (t)( φ 0 φ 1 )(α)(t)for a.e. t[a,b],

which, together with (7) and (32), yields that α(t)0 for t[a,b]. Consequently, condition (35) guarantees u0, and thus the homogeneous problem (27), (28) has only the trivial solution. □

Proof of Corollary 2.1 The validity of the corollary follows immediately from Theorem 2.1 with φ 0 =ε 0 and φ 1 =(1ε) 0 + 1 . □

Proof of Theorem 2.2 Let u be a solution to problem (27), (28). Then, in view of (9), we get

| u ( t ) | =(u)(t)sgnu(t)φ ( | u | ) (t)for a.e. t[a,b].
(36)

On the other hand, by virtue of the assumptions h 0 , h 1 P F a b , condition (28) yields

λ | u ( b ) | = | u ( a ) h 0 ( u ) + h 1 ( u ) | | u ( a ) | + h 0 ( | u | ) + h 1 ( | u | ) ,

i.e.,

| u ( a ) | λ | u ( b ) | h 0 ( | u | ) h 1 ( | u | ) =ω ( | u | ) .
(37)

Taking now the assumption φ V ˜ a b (ω) into account, we get from conditions (36) and (37) that

| u ( t ) | 0for t[a,b].

Consequently, the homogeneous problem (27), (28) has only the trivial solution. □

Proof of Theorem 2.3 Suppose that problem (27), (28) possesses a nontrivial solution u. According to conditions (11)-(13) and the assumption 0 P a b , Proposition 4.2 guarantees the validity of the inclusion

1 V ˜ a b (h).

Therefore, by virtue of the assumption 0 P a b , it follows from Definition 2.1 that u changes its sign. Put

M=max { u ( t ) : t [ a , b ] } ,m=min { u ( t ) : t [ a , b ] }
(38)

and choose t M , t m [a,b] such that

u( t M )=M,u( t m )=m.
(39)

Obviously,

M>0,m>0,
(40)

and without loss of generality, we can assume that t m < t M . Using conditions (27), (28), (12), and (13), by virtue of (38), (40), and the assumption 0 P a b , we get

( M γ ( t ) + u ( t ) ) 1 ( M γ + u ) ( t ) 0 ( M u ) ( t ) 1 ( M γ + u ) ( t ) for a.e.  t [ a , b ] ,
(41)
Mγ(a)+u(a)<h(Mγ+u)
(42)

and

( m γ ( t ) u ( t ) ) 1 ( m γ u ) ( t ) 0 ( m + u ) ( t ) 1 ( m γ u ) ( t ) for a.e.  t [ a , b ] ,
(43)
mγ(a)u(a)<h(mγu).
(44)

Hence, according to the condition 1 V ˜ a b (h), inequalities (41)-(44) yield

Mγ(t)+u(t)0,mγ(t)u(t)0for t[a,b].

However, we assume that 1 P a b , and thus, it follows from (41) and (43) that

u (t)M γ (t), u (t)m γ (t)for a.e. t[a,b].
(45)

The integration of the first inequality in (45) from t m to t M , in view of (39) and (40), implies

M+mM ( γ ( t m ) γ ( t M ) ) ,

i.e.,

0<mM ( γ ( t m ) γ ( t M ) 1 ) .
(46)

On the other hand, the integrations of the second inequality in (45) from a to t m and from t M to b, in view of (39) and (40), yield

u(a)+mm ( γ ( a ) γ ( t m ) ) ,Mu(b)m ( γ ( t M ) γ ( b ) ) .
(47)

Moreover, on account of (38) and the assumptions h 0 , h 1 P F a b , condition (28) results in

u(a)λu(b)= h 0 (u) h 1 (u)m h 0 (1)M h 1 (1).

Therefore, from (47) we get

M ( λ h 1 ( 1 ) ) +m ( 1 h 0 ( 1 ) ) m ( γ ( a ) γ ( t m ) + λ ( γ ( t M ) γ ( b ) ) ) ,

which, in view of (11) and (40), yields that

0 < M ( λ h 1 ( 1 ) ) min { 1 , 1 λ } m ( γ ( a ) γ ( t m ) + γ ( t M ) γ ( b ) ( 1 h 0 ( 1 ) ) min { 1 , 1 λ } ) .
(48)

Now, from inequalities (46) and (48), we obtain

γ(a)γ(b)>1+ ( 1 h 0 ( 1 ) ) min { 1 , 1 λ }
(49)

and

( λ h 1 ( 1 ) ) min { 1 , 1 λ } ( γ ( t m ) γ ( t M ) 1 ) × ( γ ( a ) γ ( t m ) + γ ( t M ) γ ( b ) ( 1 h 0 ( 1 ) ) min { 1 , 1 λ } ) .
(50)

In view of the inequality 4xy ( x + y ) 2 , it follows from condition (50) that

4 ( λ h 1 ( 1 ) ) min { 1 , 1 λ } ( γ ( a ) γ ( b ) 1 ( 1 h 0 ( 1 ) ) min { 1 , 1 λ } ) 2 ,

which, together with (11) and (49), contradicts (14).

The contradiction obtained proves that problem (27), (28) has only the trivial solution. □

Proof of Theorem 2.4 Suppose that problem (27), (28) possesses a nontrivial solution u. According to conditions (12), (16), and (17) and the assumptions 0 P a b and h 0 P F a b , Proposition 4.2 guarantees the validity of the inclusion

1 V ˜ a b ( h ) ,

where the functional h is defined by the formula

h (v) = def λv(b) h 1 (v)for vC ( [ a , b ] ; R ) .
(51)

Therefore, by virtue of the assumptions 0 P a b and h 0 P F a b , it follows from Definition 2.1 that u changes its sign. Define the numbers M and m by formulae (38), and choose t M , t m [a,b] such that conditions (39) hold. Obviously, (40) is satisfied, and without loss of generality, we can assume that t m < t M . Using conditions (27), (28), (12), and (17), by virtue of (38), (40), (51), and the assumptions 0 P a b and h 0 P F a b , we get relations (41), (43),

Mγ(a)+u(a) h (Mγ+u) h 0 (Mu) h (Mγ+u)
(52)

and

mγ(a)u(a) h (mγu) h 0 (m+u) h (mγu).
(53)

Hence, according to the condition 1 V ˜ a b ( h ), inequalities (41), (43), (52), and (53) yield

Mγ(t)+u(t)0,mγ(t)u(t)0for t[a,b].

However, we assume that 1 P a b , and thus, it follows from (41) and (43) that inequalities (45) hold.

Now, analogously to the proof of Theorem 2.3, relations (49) and (50) can be derived. Since assumption (16) implies λ>1, we get from (50) the inequality

4 ( 1 1 λ h 1 ( 1 ) ) ( γ ( a ) γ ( b ) 1 1 h 0 ( 1 ) λ ) 2 ,

which, together with (16) and (49), contradicts (18).

The contradiction obtained proves that problem (27), (28) has only the trivial solution. □

Proof of Theorem 2.5 Let u be a solution to problem (27), (28). We first show that each of assumptions (23), (24), or (25) ensures that u does not change its sign. Indeed, suppose that, on the contrary, u changes its sign. Define the numbers M and m by formulae (38), and choose t M , t m [a,b] such that conditions (39) hold. Obviously, (40) is satisfied, and without loss of generality, we can assume that t M < t m .

  1. (a)

    Let condition (23) hold. Then the integrations of (27) from a to t M , from t M to t m , and from t m to b, in view of (38), (39), and the assumption P a b , result in

    Mu(a)= a t M (u)(s)dsM a t M (1)(s)ds,
    (54)
M+m= t M t m (u)(s)dsm t M t m (1)(s)ds,
(55)
u(b)+m= t m b (u)(s)dsM t m b (1)(s)ds.
(56)

Hence, by virtue of (40), condition (55) implies

0<Mm ( t M t m ( 1 ) ( s ) d s 1 ) .
(57)

On the other hand, on account of (38) and the assumptions h 0 , h 1 P F a b , condition (28) yields

λu(b)u(a)= h 1 (u) h 0 (u)m h 1 (1)M h 0 (1).

Now, combining (54) and (56), we get

m ( λ h 1 ( 1 ) ) +M ( 1 h 0 ( 1 ) ) M ( a t M ( 1 ) ( s ) d s + λ t m b ( 1 ) ( s ) d s ) ,

which, on account of (20) and (40), yields

0 < m ( λ h 1 ( 1 ) ) min { 1 , 1 λ } M ( I ( 1 ) ( s ) d s ( 1 h 0 ( 1 ) ) min { 1 , 1 λ } ) ,
(58)

where I=[a, t M ][ t m ,b]. Now, conditions (57) and (58) yield

a b (1)(s)ds>1+ ( 1 h 0 ( 1 ) ) min { 1 , 1 λ }
(59)

and

( λ h 1 ( 1 ) ) min { 1 , 1 λ } ( t M t m ( 1 ) ( s ) d s 1 ) × ( I ( 1 ) ( s ) d s ( 1 h 0 ( 1 ) ) min { 1 , 1 λ } ) .
(60)

In view of the inequality 4xy ( x + y ) 2 , we get from condition (60) that

4 ( λ h 1 ( 1 ) ) min { 1 , 1 λ } ( a b ( 1 ) ( s ) d s 1 ( 1 h 0 ( 1 ) ) min { 1 , 1 λ } ) 2 ,

which, together with (20) and (59), contradicts (23).

  1. (b)

    If (24) holds then, in view of Definition 2.2, the assumption u(a)0 (resp. u(a)<0) implies u(t)0 (resp. u(t)0) for t[a,b], which contradicts (40).

  2. (c)

    If (25) holds, then, in view of Definition 2.2, the assumption u(b)0 (resp. u(b)<0) implies u(t)0 (resp. u(t)0) for t[a,b], which contradicts (40).

The contradictions obtained prove that u does not change its sign. We can assume without loss of generality, that the function u is nonnegative. Since P a b , it follows from equation (27) that

0u(a)u(t)u(b)for t[a,b].
(61)

Suppose that u(b)>0. Then, in view of (20), (61), and the assumptions h 0 , h 1 P F a b , condition (28) yields

u(a)=λu(b)+ h 0 (u) h 1 (u) ( λ h 1 ( 1 ) ) u(b)>0.

Hence, condition (61) implies

u(t)>0for t[a,b].
(62)

Put

v(t)=ru(t)γ(t)for t[a,b],

where

r=max { γ ( t ) u ( t ) : t [ a , b ] } .

According to (62), it is clear that

v(t)0for t[a,b]
(63)

and there exists t 0 [a,b] such that

v( t 0 )=0.
(64)

Taking now (27), (21), (63), and the assumption P a b into account, we obtain

v (t)(v)(t)0for a.e. t[a,b].

Therefore, on account of conditions (63) and (64), the latter relation yields

0=v(a)v(t)v(b)for t[a,b].
(65)

However, using (28), (20), (22), (65), and the assumptions h 0 , h 1 P F a b , we get the contradiction

0=v(a)>λv(b)+ h 0 (v) h 1 (v) ( λ h 1 ( 1 ) ) v(b)0.

The contradiction obtained proves that u(b)0, and thus, condition (61) implies u0. Consequently, the homogeneous problem (27), (28) has only the trivial solution. □

3 Differential equations with argument deviations

In this section, we give some corollaries of the main results for the equation with deviating arguments (5). Recall that we suppose that p,gL([a,b]; R + ) and τ,μ:[a,b][a,b] are measurable functions. The conditions stated below show that problem (5), (2) is uniquely solvable, provided that either the coefficients p and g are ‘small’ in a certain sense, or the deviations τ and μ are ‘close’ to the identities (the functional differential equation (5) is ‘close’ to the ordinary one).

3.1 Formulation of results

Theorem 2.1 implies the following.

Corollary 3.1 Let relations (11) be fulfilled, and let the functions p and τ satisfy at least one of the following conditions:

  1. (a)
    a b p(s)ds2 ( 1 h 0 ( 1 ) ) min { 1 , 1 λ } ;
  2. (b)

    0< h 0 (1)<1, τ(t)t for a.e. t[a,b], and

    esssup { t τ ( t ) p ( s ) d s : t [ a , b ] } < κ ,
    (66)

where

κ =sup { p L x ln 2 x e x ( 1 h 0 ( 1 ) ) p L ( e x 1 ) : 0 < x < ln 1 h 0 ( 1 ) } .

Let, moreover, the functions g and μ satisfy at least one of the following conditions:

  1. (A)
    a b g(s)ds< h ( 1 ) 1 λ + h 0 ( 1 ) ;
    (67)
  2. (B)

    g0 and

    esssup { μ ( t ) t g ( s ) d s : t [ a , b ] } < ξ ,
    (68)

where

ξ =sup { g L y ln y e y ( h ( 1 ) 1 ) g L ( e y 1 ) ( λ + h 0 ( 1 ) ) : 0 < y < ln λ + h 0 ( 1 ) 1 + h 1 ( 1 ) } .
(69)

Then problem (5), (2) has a unique solution.

From Theorem 2.3, we derive

Corollary 3.2 Let relations (11) be fulfilled and

β 0 (a)<h( β 0 ),
(70)
h ( 1 ) 1 λ h 1 ( 1 ) ( β 0 ( a ) h 1 ( β 2 ) λ + h 0 ( 1 ) h 1 ( β 0 ) + β 2 ( a ) ) < ω 1 (1 A 1 ),
(71)

where the number ω 1 is given by formula (15) and

A 1 = β 0 ( a ) λ + h 0 ( 1 ) h 1 ( β 0 ) ( 1 + h 1 ( β 1 ) ) + β 1 (a),
(72)
β 0 (t)=exp ( t b g ( s ) d s ) fort[a,b],
(73)
β 1 (t)= t b g(s)σ(s) ( μ ( s ) s g ( ξ ) d ξ ) exp ( t s g ( η ) d η ) dsfort[a,b],
(74)
β 2 (t)= t b g(s) ( μ ( s ) b p ( ξ ) d ξ ) exp ( t s g ( η ) d η ) ds+ t b p(s)dsfort[a,b]
(75)

and

σ(t)= 1 2 ( 1 + sgn ( t μ ( t ) ) ) for a.e.t[a,b].
(76)

Then problem (5), (2) has a unique solution.

Theorem 2.4 yields the following.

Corollary 3.3 Let relations (16) be fulfilled,

β 0 (a)<λ h 1 ( β 0 )
(77)

and

λ h 1 ( 1 ) 1 λ h 1 ( 1 ) ( β 0 ( a ) λ h 1 ( β 0 ) ( h 0 ( 1 ) + h 1 ( β 2 ) ) + β 2 ( a ) ) < ( ω 2 + h 0 ( 1 ) λ h 1 ( 1 ) ) ( 1 A 2 ) ,
(78)

where the functions β 0 , β 1 , β 2 , and σ are defined by formulae (73)-(76), the number ω 2 is given by formula (19), and

A 2 = β 0 ( a ) λ h 1 ( β 0 ) ( 1 + h 1 ( β 1 ) ) + β 1 (a).
(79)

Then problem (5), (2) has a unique solution.

Finally, we give statements concerning equation (5) with g0, i.e., the equation

u (t)=p(t)u ( τ ( t ) ) +q(t),
(80)

where pL([a,b]; R + ), qL([a,b];R), and τ:[a,b][a,b] is a measurable function.

From Theorem 2.1 we can derive the following.

Corollary 3.4 Let relations (11) be fulfilled,

0< a b p(s)ds3 ( 1 h 0 ( 1 ) ) min { 1 , 1 λ }
(81)

and

esssup { τ ( t ) t p ( s ) d s : t [ a , b ] } < ξ ,
(82)

where

ξ =sup { p L y ln 3 y e y ( h ( 1 ) 1 ) p L ( e y 1 ) ( λ + h 0 ( 1 ) ) : 0 < y < ln λ + h 0 ( 1 ) 1 + h 1 ( 1 ) } .

Then problem (80), (2) has a unique solution.

The next two statements follow from Theorem 2.5.

Corollary 3.5 Let p0, let the relations

h(1)<1, h 1 (1)<λ

be fulfilled, and let

essinf { t τ ( t ) p ( s ) d s : t [ a , b ] } > ξ ,
(83)

where

ξ =inf { p L y ln y e y ( 1 h ( 1 ) ) p L ( e y 1 ) ( 1 h 0 ( 1 ) ) : y > ln 1 h 0 ( 1 ) λ h 1 ( 1 ) } .
(84)

Let, moreover,

esssup { t τ ( t ) p ( s ) d s : t [ a , b ] } < ξ ,
(85)

where

ξ =sup { p L y ln y e y p L ( e y 1 ) : y > 0 } .
(86)

Then problem (80), (2) has a unique solution.

Corollary 3.6 Let p0, let the relations

h(1)1, h 0 (1)1, h 1 (1)<λ

be fulfilled, and let condition (85) hold, where the number ξ is defined by formula (86). Then problem (80), (2) has a unique solution.

3.2 Proofs

Proof of Corollary 3.1 Let the operators 0 and 1 be defined by the formulae

0 (v)(t) = def p(t)v ( τ ( t ) ) for a.e. t[a,b] and all vC ( [ a , b ] ; R )
(87)

and

1 (v)(t) = def g(t)v ( μ ( t ) ) for a.e. t[a,b] and all vC ( [ a , b ] ; R ) .
(88)

It is easy to verify that both conditions (a) and (b) of the corollary yield

1 2 0 V ˜ a b (h)

(see Propositions 4.3 and 4.4).

On the other hand, both conditions (A) and (B) of the corollary guarantee the validity of the inclusion

1 V ˜ a b (h)

(see Propositions 4.5 and 4.6).

Consequently, the assumptions of Corollary 2.1 are satisfied with ε= 1 2 . □

Proof of Corollary 3.2 Let the operators 0 and 1 be defined by formulae (87) and (88), respectively. According to condition (71), there exists ε>0 such that

h ( 1 ) 1 λ h 1 ( 1 ) ( β 0 ( a ) ( ε + h 1 ( β 2 ) ) λ + h 0 ( 1 ) h 1 ( β 0 ) + β 2 ( a ) ) ω 1 (1 A 1 ).
(89)

Moreover, conditions (11), (70), and (71) imply A 1 <1. Therefore, by virtue of (70) and (72), Proposition 4.7 guarantees the validity of the inclusion

1 V ˜ a b (h).

Hence, according to Remark 2.6, the problem

γ (t)=g(t)γ ( μ ( t ) ) p(t),
(90)
γ(a)=h(γ)ε
(91)

has a unique solution γ. It is clear that the function γ satisfies conditions (12) and (13). Using the inclusion 1 V ˜ a b (h), we get γ(t)0 for t[a,b], and thus, equation (90) yields

0γ(b)γ(t)γ(a)for t[a,b].
(92)

Furthermore, on account of (11), (92), and the assumptions h 0 , h 1 P F a b , condition (91) implies

λγ(b)=γ(a) h 0 (γ)+ h 1 (γ)+ε>γ(a) ( 1 h 0 ( 1 ) ) 0.

Therefore, condition (92) yields that γ(t)>0 for t[a,b].

On the other hand, γ is a solution to the equation

γ (t)=g(t)γ(t)g(t) μ ( t ) t g(s)γ ( μ ( s ) ) dsg(t) μ ( t ) t p(s)dsp(t).

Hence, in view of notations (73) and (75), the Cauchy formula implies

γ(t)=γ(b) β 0 (t)+ t b g(s) ( μ ( s ) s g ( ξ ) γ ( μ ( ξ ) ) d ξ ) exp ( t s g ( η ) d η ) + β 2 (t)

for t[a,b], whence we get

γ(t)γ(b) β 0 (t)+γ(a) β 1 (t)+ β 2 (t)for t[a,b].
(93)

Taking now conditions (92), (93) and the assumptions h 0 , h 1 P F a b into account, the relation (91) yields

γ(a)γ(b) ( λ + h 0 ( 1 ) h 1 ( β 0 ) ) γ(a) h 1 ( β 1 ) h 1 ( β 2 )ε.
(94)

Therefore, we get from (93) and (94) the inequality

γ(a)γ(a) A 1 + β 0 ( a ) ( ε + h 1 ( β 2 ) ) λ + h 0 ( 1 ) h 1 ( β 0 ) + β 2 (a).
(95)

On the other hand, by virtue of (92) and the assumptions h 0 , h 1 P F a b , condition (91) implies

γ(a)=λγ(b)+ h 0 (γ) h 1 (γ)ε< ( λ h 1 ( 1 ) ) γ(b)+γ(a) h 0 (1),

and thus,

γ(a)γ(b)< h ( 1 ) 1 λ h 1 ( 1 ) γ(a).
(96)

Now, it is clear that conditions (89), (95), and (96) guarantee the validity of inequality (14).

Consequently, the assumptions of Theorem 2.3 are satisfied. □

Proof of Corollary 3.3 Let the operators 0 and 1 be defined by formulae (87) and (88), respectively. Condition (78) implies A 2 <1. Therefore, according to (77) and (79), Proposition 4.7 guarantees the validity of the inclusion

1 V ˜ a b ( h ) ,
(97)

where the functional h is defined by formula (51). Hence, by virtue of Remark 2.6, equation (90) has a unique solution γ satisfying the boundary condition

γ(a)= h (γ) h 0 (1).
(98)

It is clear that the function γ satisfies conditions (12) and (17). Using inclusion (97), we get γ(t)0 for t[a,b], and thus, equation (90) yields the relation (92). Moreover, on account of (16), (92) and the assumption h 1 P F a b , condition (98) implies

λγ(b)=γ(a)+ h 1 (γ)+ h 0 (1)>0.

Therefore, condition (92) yields that γ(t)>0 for t[a,b].

On the other hand, γ is a solution to the equation

γ (t)=g(t)γ(t)g(t) μ ( t ) t g(s)γ ( μ ( s ) ) dsg(t) μ ( t ) t p(s)dsp(t).

Hence, in view of notations (73) and (75), the Cauchy formula implies

γ(t)=γ(b) β 0 (t)+ t b g(s) ( μ ( s ) s g ( ξ ) γ ( μ ( ξ ) ) d ξ ) exp ( t s g ( η ) d η ) + β 2 (t)

for t[a,b], whence we get relation (93). Taking now (93) and the assumption h 1 P F a b into account, condition (98) yields

γ(a)γ(b) ( λ h 1 ( β 0 ) ) γ(a) h 1 ( β 1 ) h 1 ( β 2 ) h 0 (1).
(99)

Therefore, we get from (93) and (99) the inequality

γ(a)γ(a) A 2 + β 0 ( a ) λ h 1 ( β 0 ) ( h 0 ( 1 ) + h 1 ( β 2 ) ) + β 2 (a).
(100)

On the other hand, by virtue of (92) and the assumption h 1 P F a b , condition (98) implies

γ(a)=λγ(b) h 1 (γ) h 0 (1) ( λ h 1 ( 1 ) ) γ(b) h 0 (1),

and thus,

γ(a)γ(b) λ h 1 ( 1 ) 1 λ h 1 ( 1 ) γ(a) h 0 ( 1 ) λ h 1 ( 1 ) .
(101)

Now it is clear that conditions (78), (100), and (101) guarantee the validity of inequality (18).

Consequently, the assumptions of Theorem 2.4 are satisfied. □

Proof of Corollary 3.4 Let the operator 0 be defined by formula (87), and let 1 0. It is easy to verify that conditions (81) and (82) yield

1 3 0 V ˜ a b (h), 1 3 0 V ˜ a b (h)

(see Propositions 4.3 and 4.6).

Consequently, assumptions of Corollary 2.1 are satisfied with ε= 1 3 . □

Proof of Corollary 3.5 Let the operator be defined by the formula

(v)(t) = def p(t)v ( τ ( t ) ) for a.e. t[a,b] and all vC ( [ a , b ] ; R ) .
(102)

It is clear that P a b . Moreover, condition (85) implies the validity of inclusion (24) (see Proposition 4.8).

On the other hand, according to (83) and (84), there exist y 0 >0 and ε>0 such that

y 0 ln 1 h 0 ( 1 ) + ε λ h 1 ( 1 )
(103)

and

t τ ( t ) p(s)ds p L y 0 ln y 0 e y 0 p L ( e y 0 + δ ) for a.e. t[a,b],
(104)

where

δ= ( λ h 1 ( 1 ) ) e y 0 + h 0 ( 1 ) 1 ε 1 h ( 1 ) .
(105)

Obviously, condition (103) yields δ0. Therefore, we get from (104) the relation

e y 0 p L a τ ( t ) p ( s ) d s +δ y 0 p L e y 0 p L a t p ( s ) d s for a.e. t[a,b].
(106)

Now, we put

γ(t)= e y 0 p L a t p ( s ) d s +δfor t[a,b].

Then, by virtue of (105), (106), and the assumptions h 0 , h 1 P F a b , it is easy to verify that the function γ satisfies conditions (21) and (22).

Consequently, the assumptions of Theorem 2.5 are fulfilled. □

Proof of Corollary 3.6 Let the operator be defined by formula (102). It is clear that P a b . Moreover, condition (85) implies the validity of inclusion (24) (see Proposition 4.8).

Consequently, by virtue of Remark 2.4, the assumptions of Theorem 2.5 are satisfied. □

4 On the set V ˜ a b (h)

In this section, we give some sufficient conditions guaranteeing the inclusions V ˜ a b (h), S a b (a), and S a b (b), which are stated in [22, 23]. We first formulate rather general results.

Proposition 4.1 [[22], Cor. 4.1]

Let P a b be a b-Volterra operator, and let the functional h be defined by formula (4), where λ>0 and h 0 , h 1 P F a b are such that inequalities (11) are fulfilled. If there exists a function γ C ˜ ([a,b];]0,+[) satisfying

γ (t)(γ)(t)for a.e.t[a,b],γ(a) h 0 (γ),

then V ˜ a b (h).

Proposition 4.2 [[22], Thms. 3.2 and 4.3]

Let P a b , and let the functional h be defined by formula (4), where λ>0 and h 0 , h 1 P F a b are such that inequalities (11) are fulfilled. Then V ˜ a b (h) if and only if there exists a function γ C ˜ ([a,b];]0,+[) satisfying

γ (t)(γ)(t)for a.e.t[a,b],γ(a)<h(γ).

Choosing suitable functions γ in the propositions stated above, we can derive several efficient conditions sufficient for the validity of the inclusion V ˜ a b (h). These conditions are not formulated here in detail; we present, however, some of their corollaries for ‘operators with argument deviations,’ which are used in the proofs of the results stated in Section 3.

Proposition 4.3 [[22], Cor. 5.3]

Let pL([a,b]; R + ), τ:[a,b][a,b] be a measurable function, and let the functional h be defined by formula (4), where λ>0 and h 0 , h 1 P F a b are such that inequalities (11) are fulfilled. If

a b p(s)ds ( 1 h 0 ( 1 ) ) min { 1 , 1 λ } ,

then the operator , defined by formula (102), belongs to the set V ˜ a b (h).

Proposition 4.4 [[22], Thm. 5.3(c)]

Let pL([a,b]; R + ), τ:[a,b][a,b] be a measurable function, and let the functional h be defined by formula (4), where λ>0 and h 0 , h 1 P F a b are such that the inequalities

h(1)>1,0< h 0 (1)<1

are fulfilled. Assume that τ(t)t for a.e. t[a,b], and inequality (66) holds, where

κ =sup { p L x ln x e x ( 1 h 0 ( 1 ) ) p L ( e x 1 ) : 0 < x < ln 1 h 0 ( 1 ) } .

Then the operator , defined by formula (102), belongs to the set V ˜ a b (h).

Proposition 4.5 [[22], Rem. 4.3]

Let gL([a,b]; R + ), μ:[a,b][a,b] be a measurable function, and let the functional h be defined by formula (4), where λ>0 and h 0 , h 1 P F a b are such that inequalities (11) are fulfilled. If, moreover, inequality (67) is satisfied, then the operator , defined by the formula

(v)(t) = def g(t)v ( μ ( t ) ) for a.e.t[a,b]and allvC ( [ a , b ] ; R ) ,
(107)

belongs to the set V ˜ a b (h).

Proposition 4.6 [[22], Cor. 5.2]

Let gL([a,b]; R + ), μ:[a,b][a,b] be a measurable function, and let the functional h be defined by formula (4), where λ>0 and h 0 , h 1 P F a b are such that inequalities (11) are fulfilled. If, moreover, g0 and inequality (68) is satisfied, where the number ξ is given by formula (69), then the operator , defined by formula (107), belongs to the set V ˜ a b (h).

Proposition 4.7 [[22], Thm. 5.7]

Let gL([a,b]; R + ), μ:[a,b][a,b] be a measurable function, and let the functional h be defined by formula (4), where λ>0 and h 0 , h 1 P F a b are such that inequalities (11) are fulfilled. If, moreover, inequalities (70) and

β 0 (a) ( 1 h 1 ( β 1 ) ) +h( β 0 ) β 1 (a)<h( β 0 )

are satisfied, where the functions β 0 and β 1 are defined by formulae (73), (74), and (76), then the operator , defined by formula (107), belongs to the set V ˜ a b (h).

The last statement concerns the set S a b (a).

Proposition 4.8 [[23], Thm. 1.9]

Let pL([a,b]; R + ), p0, be such that inequality (85) is satisfied, where the number ξ is defined by formula (86). Then the operator , defined by formula (102), belongs to the set S a b (a).

Author’s contributions

The author read and approved the final manuscript.

References

  1. 1.

    Azbelev NV, Maksimov VP, Rakhmatullina LF: Introduction to the Theory of Functional Differential Equations. Nauka, Moscow; 1991. (in Russian)

  2. 2.

    Hale J: Theory of Functional Differential Equations. Springer, Berlin; 1977.

  3. 3.

    Schwabik Š, Tvrdý M, Vejvoda O: Differential and Integral Equations: Boundary Value Problems and Adjoints. Academia, Prague; 1979.

  4. 4.

    Kolmanovskii V, Myshkis A: Introduction to the Theory and Applications of Functional Differential Equations. Kluwer Academic, Dordrecht; 1999.

  5. 5.

    Kiguradze I, Sokhadze Z: On the global solvability of the Cauchy problem for singular functional differential equations. Georgian Math. J. 1997, 4(4):355–373. 10.1023/A:1022994513010

  6. 6.

    Kiguradze IT, Půža B: Teoremy tipa Konti-Opiala dlja nelinejnych funkcional’nych defferentsial’nych uravnenij. Differ. Uravn. 1997, 33(2):185–194. (in Russian)

  7. 7.

    Kiguradze I, Půža B: On boundary value problems for functional differential equations. Mem. Differ. Equ. Math. Phys. 1997, 12: 106–113.

  8. 8.

    Dilnaya N, Rontó A: Multistage iterations and solvability of linear Cauchy problems. Miskolc Math. Notes 2003, 4(2):89–102.

  9. 9.

    Ronto AN:Exact solvability conditions of the Cauchy problem for systems of linear first-order functional differential equations determined by ( σ 1 , σ 2 ,, σ n ;τ)-positive operators. Ukr. Math. J. 2003, 55(11):1853–1884.

  10. 10.

    Kiguradze I, Sokhadze Z: On the uniqueness of a solution to the Cauchy problem for functional differential equations. Differ. Equ. 1995, 31(12):1947–1958.

  11. 11.

    Kiguradze I, Půža B Folia Facult. Scien. Natur. Univ. Masar. Brunensis. In Boundary Value Problems for Systems of Linear Functional Differential Equations. Masaryk University, Brno; 2003.

  12. 12.

    Hakl R, Lomtatidze A, Půža B: On a boundary value problem for first order scalar functional differential equations. Nonlinear Anal. 2003, 53(3–4):391–405. 10.1016/S0362-546X(02)00305-X

  13. 13.

    Hakl R, Lomtatidze A, Šremr J Folia Facult. Scien. Natur. Univ. Masar. Brunensis. In Some Boundary Value Problems for First Order Scalar Functional Differential Equations. Masaryk University, Brno; 2002.

  14. 14.

    Kiguradze I, Půža B: On boundary value problems for systems of linear functional differential equations. Czechoslov. Math. J. 1997, 47: 341–373. 10.1023/A:1022829931363

  15. 15.

    Hakl R, Lomtatidze A, Stavroulakis IP: On a boundary value problem for scalar linear functional differential equations. Abstr. Appl. Anal. 2004, 2004(1):45–67. 10.1155/S1085337504309061

  16. 16.

    Hakl R, Kiguradze I, Půža B: Upper and lower solutions of boundary value problems for functional differential equations and theorems on functional differential inequalities. Georgian Math. J. 2000, 7(3):489–512.

  17. 17.

    Hakl R, Lomtatidze A, Šremr J: Solvability and unique solvability of a periodic type boundary value problems for first order scalar functional differential equations. Georgian Math. J. 2002, 9(3):525–547.

  18. 18.

    Hakl R, Lomtatidze A, Šremr J: Solvability of a periodic type boundary value problem for first order scalar functional differential equations. Arch. Math. 2004, 40(1):89–109.

  19. 19.

    Hakl R, Lomtatidze A, Šremr J: On a periodic type boundary value problem for first order linear functional differential equations. Neliniini Koliv. 2002, 5(3):416–433.

  20. 20.

    Lomtatidze A, Opluštil Z, Šremr J: On a nonlocal boundary value problem for first order linear functional differential equations. Mem. Differ. Equ. Math. Phys. 2007, 41: 69–85.

  21. 21.

    Lomtatidze A, Opluštil Z, Šremr J: Solvability conditions for a nonlocal boundary value problem for linear functional differential equations. Fasc. Math. 2009, 41: 81–96.

  22. 22.

    Lomtatidze A, Opluštil Z, Šremr J: Nonpositive solutions to a certain functional differential inequality. Nonlinear Oscil. 2009, 12(4):447–591. 10.1007/s11072-010-0087-z

  23. 23.

    Hakl R, Lomtatidze A, Půža B: On nonnegative solutions of first order scalar functional differential equations. Mem. Differ. Equ. Math. Phys. 2001, 23: 51–84.

  24. 24.

    Bravyi E: A note on the Fredholm property of boundary value problems for linear functional differential equations. Mem. Differ. Equ. Math. Phys. 2000, 20: 133–135.

Download references

Acknowledgements

Published results were supported by the project Popularization of BUT R&D results and support systematic collaboration with Czech students CZ.1.07/2.3.00/35.0004 and by Grant No. FSI-S-11-3 ‘Modern methods of mathematical problem modelling in engineering.’

Author information

Correspondence to Zdeněk Opluštil.

Additional information

Competing interests

The author declares that they have no competing interests.

Rights and permissions

Open Access This article is distributed under the terms of the Creative Commons Attribution 2.0 International License (https://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Reprints and Permissions

About this article

Keywords

  • functional differential equation
  • nonlocal boundary value problem
  • unique solvability
  • differential inequality