Non-Linear First-Order Differential Boundary Problems with Multipoint and Integral Conditions

: This paper considers boundary value problem (BVP) for nonlinear ﬁrst-order differential problems with multipoint and integral boundary conditions. A suitable Green function was con-structed for the ﬁrst time in order to reduce this problem into a corresponding integral equation. So that by using the Banach contraction mapping principle (BCMP) and Schaefer’s ﬁxed point theorem (SFPT) on the integral equation, we can show that the solution of the multipoint problem exists and it is unique.


Introduction
Boundary value problems for ODEs, with special initial-boundary conditions, are intensively investigated for their many applications in physics and mathematics [1,2] in a wide range of problems from vibrations to the theory of elasticity [3]. In mathematical terms, these problems are often described by the multipoint boundary value problems [1]. This theory was mainly described in the original papers of Il'in and Moiseev [4], with further developments by several authors who contributed with fundamental results based on the Leray-Schauder Continuation Theorem and corresponding nonlinear generalizations, the degree theory, and fixed point theorem (FPT).
Many authors have studied various aspects of boundary value problems with multipoint boundary conditions for the differential equations having broad applications in several branches of physics and applied mathematics [5][6][7][8][9][10][11].
Differential equations with integral boundary conditions also have many applications in modeling and analyzing of many physical systems as blood flow problems, chemical engineering, thermoelasticity, underground water flow, population dynamics, etc. .
In this paper, an original approach based on the construction of a suitable Green function is proposed for the analysis of the multipoint BPV, so that a problem on a differential system is converted into an equivalent integral equation. Comparing with the results obtained by Multy and Sivasundaram [34], we do not use the fundamental matrix of the equation. The main advantage of our choice is that we don't require the existence of the derivative of the equation with respect to the phase coordinates. Then the uniqueness of the solution is studied for the integral equation by means of the Banach contraction mapping principle (BCMP), while the existence is also shown by using Schaefer's fixed point theorem (SFPT).
The organization of the paper is as follows: in Section 2 are given some preliminary remarks about this problem, together with some related definitions and known methods. Section 3 deals with the proof of uniqueness, while in Section 4 is given a proof of the existence by means of the fixed point theorem. Some applications are given in Section 5. Conclusion and future perspectives are discussed in Section 6.

Preliminary Remarks
Let us start by considering the following nonlinear differential system .
with multipoint boundary conditions In general the solution of (1)-(2) is characterized by the following: (1) and (2) if .
Let us now study the following problem: .
We have that Lemma 1. For y ∈ C([0, T]; R n ) the solution of the BVP (3) and (4) is unique and it is given by Proof. For any t ∈ (0, T) the solution x = x(·) fulfills where x 0 is an arbitrary constant vector. Let us chose x 0 in such a way that x(t) fulfills Equation (4). There follows which implies If we put this value into Equation (5), we get Since the equality holds, from Equation (7) we get From where, by taking t ∈ [0, t 1 ], there follows Let us write this equation in the equivalent form where E is the identity matrix. Since If we define: Equation (9) becomes, For t ∈ (t 1 , t 2 ], Equation (8) gives So that by defining .

Uniqueness of the Solution
The uniqueness of the solution of problem (1) and (2) is proven here by taking into account the following: Lemma 2. Let f ∈ C([0, T] × R n ; R n ) , then x(t) is solution of the BVP (1)-(2) iff x(t) is a solution of the following integral equation Proof. The proof is obtained similarly to Lemma 1 so that by a direct computation, we can see that the solution of Equation (11) fulfills BVP (1) and (2).
Let us now assume that:

Hypothesis 1 (H1). f : [0, T] × R n → R n is a continuous function;
Hypothesis 2 (H2). There exists a constant M ≥ 0 such that the inequality holds for each t ∈ [0, T] and all x, y ∈ R n ;

Hypothesis 3 (H3).
There exists a constant K ≥ 0 such that | f (t, x)| ≤ K for each t ∈ [0, T] and all x ∈ R n .
We can show that: Theorem 1.
[Uniqueness]. By assuming that, (H1) and (H2) holds and where S = max be an operator, defined as whose fixed points are solutions of Equations (1) and (2).
Setting max [0,T] | f (t, 0)| = M f and taking r ≥ For x ∈ B r , by using (H2), we get Let us show now that F is a contraction map for any x, y ∈ B r . Thus we write It shows that according to (12), F is a contraction map and therefore (1) and (2) admits a unique solution.

Existence of the Solution
Theorem 2.
Proof. By taking into account its definition (13), we can use the SFPT to show that there exists a fixed point for F. The multistep proof is as follows: Step 1: F is a continuous operator. In order to show this, let {x n } be a sequence such that x n → x in C([0, T]; R n ). There follows that, for t ∈ [0, T] From this we get (Fx)(t) − (Fx n )(t) → 0 as n → ∞ , which implies that F is a continuous operator.
Step 2: The operator F maps bounded sets into bounded sets in C([0, T]; R n ). In order to show this, it is enough to prove that for any η > 0 there exists a positive constant ω such that for each Step 3: Let us show now that F maps bounded sets into equicontinuous sets in C([0, T]; R n ). Let ξ 1 , ξ 2 ∈ [0, T], ξ 1 < ξ 2 , B η be a bounded set in C([0, T]; R n ) as shown in Step 2, and let x ∈ B η .
The r.h.s. of this equation tends to zero for t 2 → t 1 . As a consequence of Steps 1-3 and by taking into account the Ascoli-Arzela theorem, there follows that F : C([0, T]; R n ) → C([0, T]; R n ) is continuous.
Step 4: Let us show now that the set Therefore, ∆ is bounded and we can conclude that the operator F admits at least one fixed point. As a consequence, there exists at least one solution for the problem (1) and (2) on the interval [0, T].

Example: Analysis of the Vibrations of a Non-Homogeneous String
The existence and uniqueness of the solution for a nonlinear first-order equation with multipoint boundary conditions are given for a concrete example.
be a given differential system with the following three-point and integral boundary conditions: Evidently Obviously for this case Then for t ∈ [0, 1] we obtain and for t ∈ (1, 2] Here S ≤ 2, T = 2 and K = max{|α|, |β|} So L = TSK = 2 · 2 · max{|α|, |β|} < 1. Thus max{|α|, |β|} < 1 4 . From here, we can easily see that the given system (14)-(15) has a unique solution. The solution of system (14)-(15) involves elliptic functions; therefore, the exact solution is a quite impossible problem. However, in some special cases, it is possible to obtain the exact form of the two functions x 1 (t), x 2 (t), which also fulfill the boundary-integral conditions (15). In fact, let us compute the solution of system (14)-(15) after linearization of the trigonometric functions of (14), that is .
Moreover, if we also assume that the initial conditions are as follows the two functions x 1 (t), x 2 (t), also fulfill the three-point boundary-integral conditions (15). So that, at least in the linearized case, we have explicitly computed the solution of the three-point boundary-integral problem. At the second-order approximation .
the solution can be obtained by solving the equation ..

Conclusions
In this paper, a proof of the existence and uniqueness is proved for the solution for a class of nonlinear differential equations with some special boundary conditions. These theorems might be useful in the analysis of several physical problems arising in applied fields, such as problems with impulsive conditions, or wave propagations in non-homogeneous media. So, when the hypotheses of the theorems are fulfilled, then the solution exists and is unique.
The approach given here may be applied to the special cases, for instance, if a physical process is described in terms of a multipoint boundary and is subjected to an impulsive effect at certain points, then it can be studied by the following problem: Here 0 = t 0 < t 1 < . . . < t m−1 < t m = T; 0 < τ 1 < τ 2 < . . . < τ k < T; n(t) ∈ R n×n is a given function; l i ∈ R n×n , i = 1, 2, . . . , m are given matrices; α ∈ R n is a given vector, and