Next Article in Journal
Fixed-Point Results for Krasnoselskii, Meir–Keeler, and Boyd–Wong-Type Mappings with Applications to Dynamic Market Equilibrium
Previous Article in Journal
Neutrosophic Strongly Preopen Sets and Neutrosophic Strong Precontinuity
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

On Non-Linear Differential Systems with Mixed Boundary Conditions

Institute of Mathematics, University of Miskolc, H-3515 Miskolc, Hungary
Axioms 2024, 13(12), 866; https://doi.org/10.3390/axioms13120866
Submission received: 1 November 2024 / Revised: 4 December 2024 / Accepted: 6 December 2024 / Published: 11 December 2024
(This article belongs to the Section Mathematical Analysis)

Abstract

:
For the constructive analysis of locally Lipschitzian system of non-linear differential equations with mixed periodic and two-point non-linear boundary conditions, a numerical-analytic approach is developed, which allows one to study the solvability and construct approximations to the solution. The values of the unknown solution at the two extreme points of the given interval are considered as vector parameters whose dimension is the same as the dimension of the given differential equation. The original problem can be reduced to two auxiliary ones, with simple separable boundary conditions. To study these problems, we introduce two different types of parametrized successive approximations in analytic form. To prove the uniform convergence of these series, we use the appropriate technique to see that they form Cauchy sequences in the corresponding Banach spaces. The two parametrized limit functions and the given boundary conditions generate a system of algebraic equations of suitable dimensions, the so-called system of determining equations, which give the numerical values of the introduced unknown parameters. We prove that the system of determining equations define all possible solutions of the given boundary value problems in the domain of definition. We established also the existence of the solution based on the approximate determining system, which can always be produced in practice. The theory was presented in detail in the case of a system of differential equations consisting of two equations and having two different solutions.

1. Introduction and Subsidary Statements

For the investigation of solutions of different types of periodic and nonlinear boundary value problems for ordinary differential equations side by side with numerical methods, is often are used an approximate techniques based upon some types of successive approximations constructed in analytic form. See, e.g., [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19].
We study the following non -linear differential system with mixed, partly periodic and general two-point boundary conditions on the given compact interval
d x ( t ) d t = f 1 t , x ( t ) , y ( t ) , t a , b , x R p , y R q , f 1 R p d y ( t ) d t = f 2 t , x ( t ) , y ( t ) , t a , b , x R p , y R q , f 2 R q
x ( a ) = x ( b ) ,
B y ( a ) , y ( b ) = d .
Here we suppose that f 1 : a , b × D x × D y R p ,   f 2 : a , b × D x × D y R q are continuous functions defined on bounded sets D x R p , D y R q and B : D a y × D b y R q , d R q (the domains D x , D y , D a y and D b y will be concretized later, see (17), (19) and (21)). Moreover, the functions f 1 , f 2 , B are Lipschitzian in the following form:
f 1 ( t , x , y ) f 1 ( t , x ˜ , y ˜ ) K 1 x x ˜ + K 2 y y ˜ ,
f 2 ( t , x , y ) f 2 ( t , x ˜ , y ˜ ) K 3 x x ˜ + K 4 y y ˜
for any t a , b , x , x ˜ D x ,   y , y ˜ D y and
B ( u , v ) B ( u ˜ , v ˜ ) K 5 u u ˜ + K 6 v v ˜
for all u , u ˜ D a y , v , v ˜ D b y , where K 1 , K 2 , K 3 , K 4 , K 5 , K 6 are non-negative constant matrices respectively of dimensions p × p , p × q , q × p , q × q , q × q , q × q .
By a solution of problem (1)–(3) we understand a pair of a continuously differentiable functions x : a , b D x R p and y : a , b D y R q satisfying Equation (1) on the interval a , b and conditions (2) and (3).
Here and below, the absolute value sign and inequalities between vectors are understood componentwise. A similar convention is adopted for the operations “ · ”, “max”, “min”. The symbol I n stands for the unit matrix of dimension n, r ( K ) denotes a spectral radius of a square matrix K .
On the base of matrices K 1 , K 2 , K 3 , K 4 let us introduce the ( p + q ) × ( p + q ) dimensional square matrix
K = K 1 K 2 K 3 K 4
Moreover, we suppose that for the maximal in modulus eigenvalue of matrix
Q = 3 ( b a ) 10 K
holds
r ( Q ) < 1 .
If w R n and ρ R n is a vector with non-negative components, O ρ ( w ) stands for the componentwise ρ -neighbourhood of w :
O ρ ( w ) : = ξ R n : ξ w ρ .
Similarly, for a given bounded connected set Ω R n , we define its ρ -neighbourhood by putting
O ρ ( Ω ) : = ξ Ω O ρ ξ .
We will use the following statements.
Lemma 1 
([20]). Let Λ R k ( k 1 ) be a closed bounded set and u : a , b × Λ R n be a continuous function. Then, for an arbitrary t a , b , λ Λ the componentwise inequality
a t u ( τ , λ ) 1 b a a b u ( s , λ ) d s d τ α 1 ( t ) max ( s , λ ) a , b × Λ u ( s , λ ) min ( s , λ ) a , b × Λ u ( s , λ ) 2 ,
holds, where
α 1 ( t ) = 1 t a b a a t 1 · d s + t a b a t b 1 · d s = 2 t a 1 t a b a ,
α 1 ( t ) b a 2 , t a , b .
Lemma 2 
([21]). Let the sequence of continuous functions α m m = 0 be defined by the recurrence relation
α m + 1 ( t ) = 1 t a b a a t α m ( s ) d s + t a b a t b α m ( s ) d s , α 0 ( t ) = 1 .
Then the following estimates hold for t a , b :
α m + 1 ( t ) 10 9 3 ( b a ) 10 m α 1 ( t ) , m 0 , α m + 1 ( t ) 3 ( b a ) 10 α m ( t ) , m 1 .

2. Parametrization and Convergence of Successive Approximations

The idea that we are going to employ is based on the reduction of the given problem to a family of simple auxiliary boundary value problems [22]. Let us fix certain compact convex sets D a x = D b x R p and D a y R q , D b y R q . For given bounded connected sets define the sets
D a , b x : = ( 1 θ ) z + θ η , z D a x , η D b x , θ 0 , 1 .
D a , b y : = ( 1 θ ) γ + θ λ , γ D a y , λ D b y , θ 0 , 1 .
In practice, it is convenient to choose sets D a x , D b x , D a y   , D b y as parallelepipeds.
The following notations will be required below
δ D x × D y ( f 1 ) : = max ( t , x , y ) a , b × D x × D y f 1 ( t , x , y ) min ( t , x , y ) a , b × D x × D y f 1 ( t , x , y ) 2 , δ D x × D y ( f 2 ) : = max ( t , x , y ) a , b × D x × D y f 2 ( t , x , y ) min ( t , x , y ) a , b × D x × D y f 2 ( t , x , y ) 2
where the domain D x according to (11) is a componentwise ρ x -neighbourhood of the set D a , b x defined in (17)
D x : = O ρ x ( D a , b x )
with ρ x such that
ρ x b a 2 δ D x × D y ( f 1 ) ,
and D y is the componentwise ρ y -neighbourhood of the set D a , b y
D y : = O ρ y ( D a , b y )
with ρ y such that
ρ y b a 2 δ D x × D y ( f 2 ) .
We use an approach that was used also in [23,24] in the case of a different type of boundary value problems. Namely, we introduce the vectors of parameters
z = col ( z 1 , z 2 , , z p ) , γ = col ( γ 1 , γ 2 , , γ q ) , λ = col ( λ 1 , λ 2 , , λ q )
by formally putting
z = x ( a ) = x ( b ) , γ = y ( a ) , λ = y ( b ) .
Instead of boundary value problem (1)–(3) we will consider the following two boundary value problems with periodic and two-point linear separated parametrized conditions at a and b:
d x d t = f 1 t , x , y , t a , b ,
x ( a ) = x ( b ) = z
and
d y d t = f 2 t , x , y , t a , b ,
y ( a ) = γ , y ( b ) = λ .
We focus on the continuously differentiable solutions x : a , b D x and y : a , b D y of problem (1)–(3) with values x ( a ) D a x and y ( a ) D a y , y ( b ) D b y .
As will be seen from statements below, one can then go back to the original problem by choosing the values of the introduced parameters z , γ and λ appropriately.
Let us connect for problem (25) with the periodic parametrized boundary conditions the sequences of functions
x m + 1 ( t , z , γ , λ ) = z + a t f 1 s , x m ( s , z , γ , λ ) , y m ( s , z , γ , λ ) d s              t a b a a b f 1 s , x m ( s , z , γ , λ ) , y m ( s , z , γ , λ ) d s , t a , b , m = 0 , 1 , ,
and for the investigation of auxiliary problem (27) we introduce the sequence of functions
y m + 1 ( t , z , γ , λ ) = γ + t a b a λ γ + a t f 2 s , x m ( s , z , γ , λ ) , y m ( s , z , γ , λ ) d s          t a b a a b f 2 s , x m ( s , z , γ , λ ) , y m ( s , z , γ , λ ) d s , t a , b , m = 0 , 1 , ,
satisfying the periodic conditions (25b) and two-point parametrized conditions (26b) for arbitrary z R p , γ R q and λ R q , where
x 0 ( t , z ) = z ,
y 0 ( t , γ , λ ) = γ + t a b a λ γ
Theorem 1. 
Let conditions (4)–(9) and (19)–(22) be fulfilled. Then, for all fixed z D a x , γ D a y , λ D b y :
1. 
The functions of the sequence (27) belonging to the domain D x are continuosly differentiable on the interval a , b , and satisfy the two-point periodic boundary conditions (25b) and converges uniformly as m with respect to the domain ( t , z , γ , λ ) a , b × D a x × D a y × D b y to the limit function
x t , z , γ , λ = lim m x m ( t , z , γ , λ ) ,
which satisfies the periodic boundary conditions (25b).
2. 
The continuously differentiable functions of the sequence (28) belonging to D y for t a , b uniformly converges as m with respect to the domain ( t , z , γ , λ ) a , b × D a x × D a y × D b y to the limit function
y t , z , γ , λ = lim m y m ( t , z , γ , λ ) ,
which satisfies the boundary conditions (26b).
3. 
The limit functions x t , z , γ , λ , y t , z , γ , λ for all t a , b are the unique continuously differentiable solutions of the parametrized integral equations
x ( t ) = z + a t f 1 ( s , x ( s ) , y ( s ) ) d s t a b a a b f 1 ( s , x ( s ) , y ( s ) ) d s , y ( t ) = γ + a t f 2 ( s , x ( s ) , y ( s ) ) d s t a b a a b f 2 ( s , x ( s ) , y ( s ) ) d s + t a b a λ γ
or, equivalently, of the Cauchy problems for the modified system of integro-differential equations:
d x d t = f 1 ( t , x , y ) + Δ x ( z , γ , λ ) , x a = z , d y d t = f 2 ( t , x , y ) + Δ y ( z , γ , λ ) , y a = γ ,
where Δ x : D a x × D a y × D b y R p and Δ y : D a x × D a y × D b y R q are the mappings given by the formulas
Δ x ( z , γ , λ ) : = 1 b a a b f 1 ( s , x s , z , γ , λ , y s , z , γ , λ ) d s , Δ y ( z , γ , λ ) : = 1 b a λ γ 1 b a a b f 2 ( s , x s , z , γ , λ , y s , z , γ , λ ) d s .
4. 
The following error estimate holds:
x · , z , γ , λ x m · , z , γ , λ 10 9 α 1 ( t ) Q m ( I p + q Q ) 1 δ D x × D y ( f 1 ) δ D x × D y ( f 2 ) 1 p ,
y · , z , γ , λ y m · , z , γ , λ 10 9 α 1 ( t ) Q m ( I p + q Q ) 1 δ D x × D y ( f 1 ) δ D x × D y ( f 2 ) p + 1 p + q ,
where δ D x × D y ( f 1 ) and δ D x × D y ( f 2 ) are defined in (18), the matrix Q has form (8) and · i j denotes the components i , i + 1 , , j of the vector in the brackets · .
Proof. 
We will prove that for z D a x , γ D a y , λ D b y and t a , b the values of the functions (27) and (28) belong to the domains D x and D y respectively and they create the Cauchy sequences in the Banach spaces C a , b , R p , and C a , b , R q respectively. Indeed, using the estimate (12) of Lemma 1, relations (13), (14), (27) for m = 0 , t a , b imply that
x 1 ( t , z , γ , λ ) x 0 t , z               α 1 ( t ) max ( t , x , y ) a , b × D x × D y f 1 ( t , x 0 ( t , z ) , y 0 ( t , γ , λ ) ) 2                 min ( t , x , y ) a , b × D x × D y f 1 ( t , x 0 ( t , z ) , y 0 ( t , γ , λ ) ) 2                                α 1 ( t ) δ D x × D y ( f 1 ) b a 2 δ D x × D y ( f 1 ) ,    
which means, according to (20), that x 1 t , z , γ , λ D x , whenever t , z , γ , λ a , b × D a x × D a y × D b y .
We can find similarly to (38) on the base of (28) that
y 1 ( t , z , γ , λ ) y 0 ( t , z , γ , λ ) α 1 ( t ) δ D x × D y ( f 2 ) b a 2 δ D x × D y ( f 2 ) ,
which, according to (22), means that y 1 ( t , γ , λ ) D y .
By induction, we obtain
x m t , z , γ , λ x 0 t , z , γ , λ α 1 ( t ) δ a , b × D x × D y ( f 1 ) b a 2 δ a , b × D x × D y ( f 1 )
and
y m ( t , z , γ , λ ) y 0 ( t , z , γ , λ ) α 1 ( t ) δ D x × D y ( f 2 ) b a 2 δ D x × D y ( f 2 )
which means according to (20) and (22) that all the values of functions (27) are contained in domain D x for all m = 1 , 2 , 3 , and t , z , γ , λ a , b × D a x × D a y × D b y , and all the values of functions (28) belong to D y .
Estimates (38) and (39) can be rewritten in componentwise vector form
x 1 t , z , γ , λ x 0 t , z , γ , λ y 1 ( t , z , γ , λ ) y 0 ( t , γ , λ ) α 1 ( t ) δ D x × D y ( f 1 ) α 1 ( t ) δ D x × D y ( f 2 ) .
Consider now the difference of functions
x 2 t , z , γ , λ x 1 t , z , γ , λ = = a t f 1 s , x 1 s , z , γ , λ , y 1 s , z , γ , λ f 1 s , x 0 s , z , γ , λ , y 0 s , z , γ , λ d s t a b a a b f 1 s , x 1 s , z , γ , λ , y 1 s , z , γ , λ f 1 s , x 0 s , z , γ , λ , y 0 s , z , γ , λ d s
and
y 2 ( t , z , γ , λ ) y 1 ( t , z , γ , λ ) = = a t f 2 s , x 1 s , z , γ , λ , y 1 s , z , γ , λ f 2 s , x 0 s , z , γ , λ , y 0 s , z , γ , λ d s t a b a a b f 2 s , x 1 s , z , γ , λ , y 1 s , z , γ , λ f 2 s , x 0 s , z , γ , λ , y 0 s , z , γ , λ d s .
In view of Lemma 1, Lemma 2 and conditions (4), (5) and estimates (40) from (41) and (42) we obtain
x 2 t , z , γ , λ x 1 t , z , γ , λ | a t K 1 x 1 s , z , γ , λ x 0 s , z , γ , λ + K 2 y 1 s , z , γ , λ y 0 s , s , γ , λ d s t a b a a b K 1 x 1 s , z , γ , λ x 0 s , z , γ , λ + K 2 y 1 s , z , γ , λ y 0 s , z , γ , λ d s K 1 α 2 ( t ) δ D x × D y ( f 1 ) + K 2 α 2 ( t ) δ a , b × D x × D y ( f 2 ) ,
y 2 t , z , γ , λ y 1 t , z , γ , λ | a t K 3 x 1 s , z , γ , λ x 0 s , z , γ , λ + K 4 y 1 s , z , γ , λ y 0 s , z , γ , λ d s t a b a a b K 3 x 1 s , z , γ , λ x 0 s , z , γ , λ + K 4 y 1 s , z , γ , λ y 0 s , z , γ , λ d s K 3 α 2 ( t ) δ D x × D y ( f 1 ) + K 4 α 2 ( t ) δ a , b × D x × D y ( f 2 ) .
Let us rewrite estimates (43) and (44) in the matrix-vector form
x 2 t , z , γ , λ x 1 t , z , γ , λ K 1 K 2 K 3 K 4 · α 2 ( t ) δ D x × D y ( f 1 ) α 2 ( t ) δ D x × D y ( f 2 ) , y 2 t , z , γ , λ y 1 t , z , γ , λ K 1 K 2 K 3 K 4 · α 2 ( t ) δ D x × D y ( f 1 ) α 2 ( t ) δ D x × D y ( f 2 ) .
Direct computations give
K 2 = K 1 2 + K 2 K 3 K 1 K 2 + K 2 2 K 3 K 1 + K 4 K 3 K 3 K 2 + K 4 2 ,
x 3 t , z , γ , λ x 2 t , z , γ , λ y 3 ( t , z , γ , λ ) y 2 ( t , z , γ , λ ) K 1 2 α 3 ( t ) δ D x × D y ( f 1 ) + K 1 K 2 α 3 ( t ) δ D x × D y ( f 2 ) + K 3 K 1 α 3 ( t ) δ D x × D y ( f 1 ) + K 3 K 2 α 3 ( t ) δ D x × D y ( f 2 ) + + K 2 K 3 α 3 ( t ) δ D x × D y ( f 1 ) + K 2 K 4 α 3 ( t ) δ D x × D y ( f 2 ) + K 4 K 3 α 3 ( t ) δ D x × D y ( f 1 ) + K 4 2 α 3 ( t ) δ D x × D y ( f 2 ) ,   
and
x 3 t , z , γ , λ x 2 t , z , γ , λ K 1 K 2 K 3 K 4 2 α 3 ( t ) δ D x × D y ( f 1 ) α 3 ( t ) δ D x × D y ( f 2 ) 1 p , y 3 t , z , γ , λ y 2 t , z , γ , λ K 1 K 2 K 3 K 4 2 α 3 ( t ) δ D x × D y ( f 1 ) α 3 ( t ) δ D x × D y ( f 2 ) p + 1 p + q ,
where · 1 p and · p + 1 p + q denotes the 1 , 2 , , p and p + 1 , , p + q components of the vector in · .
By induction using estimate (46), we can establish that
x m + 1 t , z , γ , λ x m t , z , γ , λ K 1 K 2 K 3 K 4 m α m + 1 ( t ) δ D x × D y ( f 1 ) α m + 1 ( t ) δ D x × D y ( f 2 ) 1 p ,
y m + 1 t , z , γ , λ y m t , z , γ , λ K 1 K 2 K 3 K 4 m α m + 1 ( t ) δ D x × D y ( f 1 ) α m + 1 ( t ) δ D x × D y ( f 2 ) p + 1 p + q .
Using estimate (16) of Lemma 2, from (47) we obtain
x m + 1 t , z , γ , λ x m t , z , γ , λ 10 9 α 1 ( t ) Q m δ D x × D y ( f 1 ) Q m δ D x × D y ( f 2 ) 1 p , y m + 1 t , z , γ , λ y m t , z , γ , λ 10 9 α 1 ( t ) Q m δ D x × D y ( f 1 ) Q m δ D x × D y ( f 2 ) p + 1 p + q .
Let us set
r m ( x ) = x m ( t , z , γ , λ ) x m 1 ( t , z , γ , λ ) , r m ( y ) = y m ( t , z , γ , λ ) y m 1 ( t , z , γ , λ ) .
Therefore, in view of (48) from inequalities
x m + j ( t , z , γ , λ ) x m ( t , z , γ , λ ) x m + j ( t , z , γ , λ ) x m + j 1 ( t , z , γ , λ ) + x m + j 1 ( t , z , γ , λ ) x m + j 2 ( t , z , γ , λ ) + + x m + 1 ( t , z , γ , λ ) x m ( t , z , γ , λ ) , y m + j ( t , z , γ , λ ) y m ( t , z , γ , λ ) y m + j ( t , z , γ , λ ) y m + j 1 ( t , z , γ , λ ) + y m + j 1 ( t , z , γ , λ ) y m + j 2 ( t , z , γ , λ ) + + y m + 1 ( t , z , γ , λ ) y m ( t , z , γ , λ )
on the base of (9) we have
x m + j t , z , γ , λ x m t , z , γ , λ                           i = 1 j r m + i ( x ) = 10 9 α 1 ( t ) i = 1 j Q m + i δ D x × D y ( f 1 ) δ D x × D y ( f 2 ) 1 p       10 9 α 1 ( t ) Q m i = 0 j 1 Q i δ D x × D y ( f 1 ) δ D x × D y ( f 2 ) 1 p                    10 9 α 1 ( t ) Q m ( I p + q Q ) 1 δ D x × D y ( f 1 ) δ D x × D y ( f 2 ) 1 p ,
and
y m + j t , z , γ , λ y m t , z , γ , λ                           i = 1 j r m + i ( y ) = 10 9 α 1 ( t ) i = 1 j Q m + i δ D x × D y ( f 1 ) δ D x × D y ( f 2 ) p + 1 p + q       10 9 α 1 ( t ) Q m i = 0 j 1 Q i δ D x × D y ( f 1 ) δ D x × D y ( f 2 ) p + 1 p + q                    10 9 α 1 ( t ) Q m ( I p + q Q ) 1 δ D x × D y ( f 1 ) δ D x × D y ( f 2 ) p + 1 p + q
Since, due to (9), the maximum eigenvalue of the matrix Q is less than 1, we have
i = 0 j 1 Q i I p + q Q 1 , lim m Q m = 0 p + q .
Therefore, we conclude from (49) and (50) that, according to Cauchy convergence test, the sequence x m t , z , γ , λ m = 0 of the form (27) uniformly converges in the domain t , z , γ , λ a , b × D a x × D a y × D b y to the limit function x t , z , γ , λ and the sequence y m t , z , γ , λ m = 0 of the form (28) uniformly converges in the domain t , z , γ , λ a , b × D a x × D a y × D b y to the limit function y t , z , γ , λ . Since all functions of the sequences (27) and (28) satisfy correspondingly the periodic parametrized conditions (25b) and conditions (26b) for all values of the introduced parameters z D a x , γ D a y , λ D b y the limit functions x · , z , γ , λ , y · , z , γ , λ also satisfies these conditions. Passing to the limit as m in equalities (27), (28) we show that the limit function satisfies both the integral Equation (33) and the Cauchy problem (34).
Passing to the limit as j in (49) we get the estimates (36). □

3. Connection of the Limit Functions x ( t , z , γ , η ) , y ( t , z , γ , η ) to the Solution of the Original Problem

Theorem 2. 
Under the assumptions of Theorem 1, the limit functions
x ( t , z , γ , λ ) = lim m x m ( t , z , γ , λ ) , y ( t , z , γ , λ ) = lim m y m ( t , z , γ , λ ) ,
of the sequences (27), (28) are the solution of the non-linear boundary value problem (1)–(3) if and only if the parameters z , γ , λ from (24) satisfy the system of p + 2 q algebraic or transcendental equations
Δ x ( z , γ , λ ) : = a b f 1 ( s , x s , z , γ , λ , y s , γ , λ ) d s = 0 ,
Δ y ( z , γ , λ ) : = λ γ a b f 2 ( s , x s , z , γ , λ , y s , z , γ , λ ) d s = 0 ,
Λ ( z , γ , λ ) : = B ( y a , γ , λ , y b , γ , λ ) d = 0 .
Proof. 
The proof can be carried out similarly as in [21,22]. □
Remark 1. 
The system of Equations (52)–(54) is usually referred to as a determining equations. In such a manner, the original infinite-dimensional problem (1)–(3) is reduced to a system of p + 2 q numerical equations.
The method thus consists of two parts, namely, the analytic part, when the integral Equations (33) are dealt with by using the method of successsive approximations (27), (28) and the numerical one, which consists in finding values of the p + 2 q unknown parameters from Equations (52)–(54).
The next statement proves that the system of determining Equations (52)–(54) defines all possible solutions of the original non-linear boundary value problem (1)–(3) having values in domains D x × D y .
Theorem 3. 
Let the assumptions of Theorem 1 hold. Furthermore, assume there exists some triplet of vectors z 0 , γ 0 , λ 0 D a x × D a y × D b y which sytisfy the system of determining Equations (52)–(54).
Then:
1. 
The non-linear boundary value problem (1)–(3) has a solution x 0 ( · ) , y 0 ( · ) such that
x 0 ( a ) = z 0 , y 0 ( a ) = γ 0 , y 0 ( b ) = λ 0 .
Moreover, this solution is given by the limit function of the sequence (27) and (28):
x 0 ( · ) = x · , z 0 , γ 0 , λ 0 = lim m x m ( · , z 0 , γ 0 , λ 0 ) , t a , b , y 0 ( · ) = y · , z 0 , γ 0 , λ 0 = lim m y m ( · , z 0 , γ 0 , λ 0 ) , t a , b .
2. 
If the non-linear boundary value problem (1)–(3) has a solution x 0 ( · ) , y 0 ( · ) then the system of determining Equations (52)–(54) is satisfied with
z = x 0 ( a ) , γ = y 0 a , λ = y 0 b .
Proof. 
The proof can be carried out similarly as in [21,22]. □

4. Solvability Analysis Based on the Approximate Determining System

Although Theorem 2 provides a theoretical answer to the question on the construction of a solution of the original non-linear boundary value problem (1)–(3), its application faces certain difficulties due to the fact that the explicit form of the limit functions x · , z , γ , λ , y · , z , γ , λ and consequently the explicit form of the functions
Δ x : D a x × D a y × D b y , Δ y : D a x × D a y × D b y R q , Λ : D a y × D b y R q
in (52)–(54) is usually unknown. This complication can be overcome by using the so-called approximate determining equations:
Δ x , m ( z , γ , λ ) : = a b f 1 ( s , x m s , z , γ , λ , y m s , γ , λ ) d s = 0 ,
Δ y , m ( z , γ , λ ) : = λ γ a b f 2 ( s , x m s , z , γ , λ , y m s , z , γ , λ ) d s = 0 ,
Λ m ( z , γ , λ ) : = B ( y m a , z , γ , λ , y m b , z , γ , λ ) d = 0 .
for a fixed m .
Lemma 3. 
Under the assumptions of Theorem 1 for the exact and approximate determining functions defined by (52)–(54) and (58)–(60) for any z D a x , η D a y , λ D b y , t a , b and m 1 , the following estimates hold:
Δ x z , γ , λ Δ x , m z , γ , λ                      10 27 ( b a ) 2 K 1 Q m ( I p + q Q ) 1 δ D x × D y ( f 1 ) δ D x × D y ( f 2 ) 1 p +               + 10 27 ( b a ) 2 K 2 Q m ( I p + q Q ) 1 δ D x × D y ( f 1 ) δ D x × D y ( f 2 ) p + 1 p + q ,
Δ y z , γ , λ Δ y , m z , γ , λ                      10 27 ( b a ) 2 K 3 Q m ( I p + q Q ) 1 δ D x × D y ( f 1 ) δ D x × D y ( f 2 ) 1 p +               + 10 27 ( b a ) 2 K 4 Q m ( I p + q Q ) 1 δ D x × D y ( f 2 ) δ D x × D y ( f 2 ) p + 1 p + q ,
B ( y ( a ) , y ( b ) ) B ( y m ( a ) , y m ( b ) )                      K 5 Q m ( I p + q Q ) 1 δ D x × D y ( f 1 ) δ D x × D y ( f 2 ) p + 1 p + q +               + K 6 Q m ( I p + q Q ) 1 δ D x × D y ( f 2 ) δ D x × D y ( f 2 ) p + 1 p + q .
Proof. 
Let us fix an arbitrary z , γ , λ D a x × D a y × D b y . Using the Lipschitz conditions (4)–(6) and (52)–(54), (58)–(60), estimates (36) and (37), we have
Δ x z , γ , λ Δ x , m z , γ , λ Δ y z , γ , λ Δ y , m z , η , λ B ( y ( a ) , y ( b ) ) B m ( y m ( a ) , y m ( b ) ) a b K 1 x s , z , γ , λ x m s , z , γ , λ + K 2 y s , z , γ , λ y m · , z , γ , λ d s a b K 3 x s , z , γ , λ x m s , z , γ , λ + K 4 y s , z , γ , λ y m s , z , γ , λ d s K 5 y ( a , z , γ , λ ) y m ( a , z , γ , λ ) + K 6 y ( b , z , γ , λ ) y m ( b , z , γ , λ )
and taking into account that
a b α 1 ( t ) d t = ( b a ) 2 3 , α 1 ( t ) b a 2 ,
we get from (64) just (61)–(63). □
Based on both exact and approximate determining systems (52)–(54) and (58)–(60) let us introduce the mappings H : D a x × D a y × D b y R P + 2 q and H m : D a x × D a y × D b y R P + 2 q by setting
H ( z , γ , λ ) : = a b f 1 ( s , x s , z , γ , λ , y s , z , γ , λ ) d s λ γ a b f 2 ( s , x s , z , γ , λ , y s , z , γ , λ ) d s B ( y a , γ , λ , y b , γ , λ ) d ,
H m ( z , γ , λ ) : = a b f 1 ( s , x m s , z , γ , λ , y m s , z , γ , λ ) d s λ γ a b f 2 ( s , x m s , z , γ , λ , y m s , z , γ , λ ) d s B ( y m a , γ , λ , y m b , γ , λ ) d ,
where z , γ , λ D a x × D a y × D b y .
We see from Theorem 2 that the critical points of the vector field H of the form (65) determine solutions of the non-linear problem (1)–(3). The next statement establishes a similar result based upon properties of vector field H m explicity known from (66).
Theorem 4. 
Let the assumptions of Theorem 1 hold. Moreover, one can specify an m 1 and a set
Ω : = D 1 × D 2 × D 3 R p + 2 q ,
where D 1 D a x , D 2 D a y , D 3 D b y are certain bounded open sets such that the mapping H m satisfies the relations
a b f 1 ( s , x m s , z , γ , λ , y m s , z , γ , λ ) d s Ω                      10 27 ( b a ) 2 K 1 Q m ( I p + q Q ) 1 δ D x × D y ( f 1 ) δ D x × D y ( f 2 ) 1 p +               + 10 27 ( b a ) 2 K 2 Q m ( I p + q Q ) 1 δ D x × D y ( f 1 ) δ D x × D y ( f 2 ) p + 1 p + q ,
a b f 2 ( s , x m s , z , γ , λ , y m s , z , γ , λ ) d s Ω                      10 27 ( b a ) 2 K 3 Q m ( I p + q Q ) 1 δ D x × D y ( f 1 ) δ D x × D y ( f 2 ) 1 p +               + 10 27 ( b a ) 2 K 4 Q m ( I p + q Q ) 1 δ D x × D y ( f 1 ) δ D x × D y ( f 2 ) p + 1 p + q ,
and
B ( x m a , z , γ , λ , x m b , z , γ , λ ) ) d Ω                       5 9 ( b a ) K 5 Q m ( I p + q Q ) 1 δ D x × D y ( f 1 ) δ D x × D y ( f 2 ) 1 p + + 5 9 ( b a ) K 6 Q m ( I p + q Q ) 1 δ D x × D y ( f 1 ) δ D x × D y ( f 2 ) p + 1 p + q
on the boundary Ω of the set Ω. If, in addition
deg H m , Ω , 0 0 ,
then there exists ( z * , γ * , λ * ) D 1 × D 2 × D 3 for which the function
x * ( · ) : = x · , z * , γ * , λ * , y * ( · ) : = y · , z * , γ * , λ *
is a solution of the non-linear boundary value problem (1)–(3).
In (67) the binary relation Γ is defined in [21] as a kind of strict inequality for vector functions and it means that, at every point on the boundary Ω , at least one of the components of the vector H m z , η is greater than the corresponding component of the vector at the right-hand side. The degree in (68) is the Brouwer degree because all the vectors fields are finite-dimensional. Likewise, all the terms in the right-hand side of (67) are computed explicitly e.g., by using computer algebra systems.
Proof. 
The proof can be carried out similarly as in Theorem 4 from [23]. □

5. Conclusions

The method is numerical-analytic in the sense that its realization consists of two stages. First we give an explicit construction of certain parametrized approximations in analytic form, then a system of non-linear algebraic equations is obtained for the introduced parameters, which can be solved numerically. The values of the unknown solution at the two extreme points of the given interval are considered as vector parameters whose dimension is the same as the dimension of the given differential equation. The original problem can be reduced to two auxiliary ones, with simple separable boundary conditions. To study these problems, we introduce two different types of successive approximations in analytic form. It can be noticed that in order to solve the given mixed boundary value problem, it was necessary to introduce two parametrized sequence of functions x m ( z , η , λ ) , y m ( z , η , λ ) in different ways, and to prove their uniform convergence. Theorem 3 proves that our method determines all possible approximate solutions in the given domain. Theorem 4 gives a new constructive answer for the existence of an exact solution based on a certain approximation. In the example, we determine two distinct solutions. The dimension of the determining algebraic system of equations is p + 2 q , where p is the number of periodic and q is the number of non-periodic components.
The theoretical basis for robustness, as stated in Remark 1, is that, in our approach, an infinite-dimensional original boundary value problem (1)–(3) is reduced to a finite dimensional system of p + 2 q algebraic or transcendental equations. The method thus consists of two parts, namely, the analytic part, when the integral Equation (33) are dealt with by using the method of successive approximations (27), (28) and the numerical one, which consists in finding values of the p + 2 q unknown parameters from Equations (52)–(54) (in practice, from systems (58)–(60)).
To solve nonlinear systems of Equations (58)–(60), efficient methods available in symbolic computation packages can be used. We have used Maple 14. The reliability of our approach is also facilitated by its constructiveness, because all the theoretical conditions in the theorems can be verified in practice. In the example, we have carried out this check in detail.
Our parameterization technique can be used with conditions of a general form, for example with integral restrictions. Method can be easily applied also to examine the solutions of differential equations with argument deviations. The determination of successive approximations can be complicated by integration. To eliminate this, the polynomial approximation proposed in our paper [24] can be used baseed on parameterized Lagrangian polynomials. We note that, in the example, we have already applied this.

6. Example

Let us apply the approach desribed above to the system of differential equations
d x ( t ) d t = f 1 ( t , x , y ) = x 2 y 2 + 1 36 sin 2 ( 2 π t ) + t 4 4 1 36 π 3 sin ( 2 π t ) , t 0 , 1 ,
d y ( t ) d t = f 2 ( t , x , y ) = x 2 y + t t 2 72 cos 2 ( 2 π t ) ,
considered under periodic and two-point non-linear boundary conditions
x ( a ) = x ( b ) , y 2 ( a ) + y 2 ( b ) = 1 4 .
Following (23), (24), introduce the parameters z , γ , λ .
Let us consider the following choice of subsets, where one looks for the values x ( a ) = x ( b ) = z and y ( a ) = γ , y ( b ) = λ :   D a x = D b x R p and D a y R q , D b y R q
D a x = D b x = { x : 0.3 x 0.2 } , D a y = D b y = { y : 0.6 y 0.7 } .
This choice of these sets is motivated by the fact that the zeroth approximate determining system (i.e., (58)–(60) with m = 0 ) has roots lying in these sets (73), see the second row in Table 1. Recall that, in order to obtain it, only functions (29) and (30) are used, and no iteration is yet carried out. We see that this piecewise linear function provides quite reasonable approximate values of the parameters. In this case, according to (17), we have
D a , b x = D a x = D b x , D a , b y = D a y = D b y .
For ρ x and ρ y involved in (20) and (22), we choose the vectors
ρ x = 0.3 , ρ y = 0.2 .
Then, in view of (73)–(75), the sets (19), (21) take the form
D x = x : 0.6 x 0.5 , D y = y : 0.8 y 0.9 .
We can compute the Lipschitz constants in (4)–(6) for the domains (76)
K 1 = 1.2 , K 2 = 1.8 , K 3 = 1.1 , K 4 = 0.36 , K 5 = 1.4 , K 6 = 1.4 .
Therefore direct computations give
K = 1.2 1.8 1.1 0.36 , Q = 3 ( b a ) 10 1.2 1.8 1.1 0.36 = 0.36 0.54 0.33 0.108 , r ( Q ) = 0.6745 , b a 2 δ D x × D y ( f 1 ) = 0.2925 , b a 2 δ D x × D y ( f 2 ) = 0.153 ,
and hence inequalities (20) and (22) hold.
We thus see that all the conditions of Theorem 1 are fulfilled, and the sequences of functions (27) and (28) for this example are uniformly convergent.
In the case of the given non-linear boundary value problems it is more appropriate to use a scheme with polynomial interpolation [24], when instead of (27), (28) we use the sequence { x m + 1 N + 1 ( t , z , γ , λ ) : m = 0 , 1 , 2 , } of vector polynomials of degree ( N + 1 ) and the sequence y m + 1 N + 1 ( t , z , γ , λ ) : m = 0 , 1 , 2 , of vector polynomials of degree ( N + 1 ) on a , b
     x m + 1 N + 1 ( t , z , γ , λ ) : =        = a m + 1 , 0 ( z , γ , λ ) + a m + 1 , 1 ( z , γ , λ ) t + a m + 1 , 2 ( z , γ , λ ) t 2 + + a m + 1 , N + 1 ( z , γ , λ ) t N + 1 =          = z + a t A m , 0 ( z , γ , λ ) + A m , 1 ( z , γ , λ ) t + A m , 2 ( z , γ , λ ) t 2 + + A m , N ( z , γ , λ ) t N d t                 t a b a a b A m , 0 ( z , γ , λ ) + A m , 1 ( z , γ , λ ) t + A m , 2 ( z , γ , λ ) t 2 + + A m , N ( z , γ , λ ) t N d t ,
where
A m , 0 ( z , γ , λ ) + A m , 1 ( z , γ , λ ) t + A m , 2 ( z , γ , λ ) t 2 + + A m , N ( z , γ , λ ) t N
is the Lagrange interpolation polynomial of degree N which corresponds to the term f 1 t , x m N + 1 ( t , z , γ , λ ) , y m N + 1 ( t , z , γ , λ ) in (70) over the Chebyshev nodes translated from 1 , 1 to the interval a , b . Similarly,
       y m + 1 N + 1 ( t , z , γ , λ ) : =        = b m + 1 , 0 ( z , γ , λ ) + b m + 1 , 1 ( z , γ , λ ) t + b m + 1 , 2 ( z , γ , λ ) t 2 + + b m + 1 , N + 1 ( z , γ , λ ) t N + 1 =          = z + a t B m , 0 ( z , γ , λ ) + B m , 1 ( z , γ , λ ) t + B m , 2 ( z , γ , λ ) t 2 + + B m , N ( z , γ , λ ) t N d t            t a b a B m , 0 ( z , γ , λ ) + B A m , 1 ( z , γ , λ ) t + B m , 2 ( z , γ , λ ) t 2 + + B m , N ( z , γ , λ ) t N d t +                                            + t a b a λ γ , t a , b , m = 0 , 1 , 2 , ,
where
B m , 0 ( z , γ , λ ) + B m , 1 ( z , γ , λ ) t + B m , 2 ( z , γ , λ ) t 2 + + B m , N ( z , γ , λ ) t N
is the Lagrange interpolation Chebyshev polynomials of degree N on the Chebyshev nodes translated from 1 , 1 to the interval a , b , corresponding to the Nemytskii operator defined by the term f 2 t , x m N + 1 ( t , z , γ , λ ) , y m N + 1 ( t , z , γ , λ ) in (71). Note that the coefficients of the interpolation polynomials depend on the parameters z, γ , λ .
The approximate determining system (58)–(60) takes the following form
Δ m , x ( z , γ , λ ) = a b [ A m , 0 ( z , γ , λ ) + A m , 1 ( z , γ , λ ) t + A m , 2 ( z , γ , λ ) t 2 + + A m , N ( z , γ , λ ) t N ] d t = 0 , Δ m , y ( z , γ , λ ) = λ γ a b [ B m , 0 ( z , γ , λ ) + B m , 1 ( z , γ , λ ) t + B m , 2 ( z , γ , λ ) t 2 + + B m , N ( z , γ , λ ) t N ] d t = 0 , Λ m ( z , γ , λ ) = B y m + 1 N + 1 ( a , z , γ , λ ) , y m + 1 N + 1 ( b , z , γ , λ ) d = 0 .
The justification of this polynomial approach can be found in [24].

6.1. Approximations for the Exact Solution

One can check that the pair of functions x * ( t ) = 1 6 cos ( 2 π t ) , y * ( t ) = t 2 2 is one of the solutions of the given boundary value problem. Using (78), (79) and applying Maple 14 for different values of m to implement the approximations x m ( t , z , γ , λ ) , y m ( t , z , γ , λ ) and solving the approximate determining system (80), we find the values of introduced parameters with the degree of the approximating polynomials N = 10 and N = 6 . These values are presented in Table 1 for different m. The graphs of approximations and residuals are shown on Figure 1 and Figure 2.

6.2. Approximations for the Second Solution

Carrying out computations, we see that the approximate determining system (80) along with the solution given in Table 1, has another solution in set (73), which is presented in Table 2.
With the help of the Maple program package, the second solution is obtained at the 9th iteration in the form of the 11th degree polynomials below
x 9 ( t ) = 1.148257431 t 11 5.316404582 t 10 + 10.31079622 t 9 17.33315601 t 8 + 31.52751918 t 7 31.14971134 t 6 + 3.061277190 t 5 + 10.47074659 t 4 + 0.8178200919 t 3 3.269716833 t 2 0.2674279358 t 0.09261226771 , y 9 ( t ) = 7.680451214 t 11 42.18446611 t 10 + 94.98737130 t 9 111.8199171 t 8 + 73.13069674 t 7 27.16587171 t 6 + 6.726379104 t 5 1.362951853 t 4 0.01490299810 t 3 + 0.4838387702 t 2 0.004241974928 t 0.4982447170 .
Figure 3 shows that the two solutions found look identical in appearance along the x and y coordinates, but they are shifted vertically. Figure 4 shows the residual.

Funding

This research received no external funding.

Data Availability Statement

Data is contained within the article.

Conflicts of Interest

The author declares no conflicts of interest.

References

  1. Samoilenko, A.M. Numerical-analytic method for the investigation of periodic systems of ordinary differential equations. I. Ukr. Math. Zhurnal 1965, 17, 82–93. [Google Scholar]
  2. Samoilenko, A.M. Numerical-analytic method for the investigation of periodic systems of ordinary differential equations. II. Ukr. Math. Zhurnal 1966, 18, 50–59. [Google Scholar]
  3. Samoilenko, A.M. On periodic solutions of nonlinear second -order equations. Differ. Uravn. 1967, 3, 1903–1913. [Google Scholar]
  4. Rontó, A.; Rontóová, N. On constuction of solutions of linear differential systems with argument deviations of mixed type. Symmetry 2019, 12, 1740. [Google Scholar] [CrossRef]
  5. Samoilenko, A.M.; Laptinsky, V.N. On estimates of periodic solutions of differential equations. Dokl. Akad. Nauk. Ukr. SSR Ser. A 1982, 30–32. (In Russian) [Google Scholar]
  6. Samoilenko, A.M.; Ronto, V.A. Numerical-analytic method for solving boundary-value problems for ordinary differential equations. Ukr. Math. J. 1981, 33, 356–362. [Google Scholar] [CrossRef]
  7. Evkhuta, N.A.; Zabreiko, P.P. On the convergence of the Samoilenko method of successive approximations for finding periodic solutions. Dokl. Akad. Nauk. Belarus. SSR 1985, 29, 15–18. (In Russian) [Google Scholar]
  8. Evkhuta, N.A.; Zabreiko, P.P. On the Samoilenko method for finding periodic solutions of quasilinear differential equations in a Banach space. Ukr. Math. Zhurnal 1985, 37, 162–168. (In Russian) [Google Scholar]
  9. Evkhuta, N.A.; Zabreiko, P.P. The Poincare method and Samoilenko method for the construction of periodic solutions to ordinary differential equations. Math. Nachrichten 1991, 153, 85–99. [Google Scholar] [CrossRef]
  10. Pylypenko, V.; Rontó, A. A singular problem for functional differential equations with decreasing non-linearities. Math. Methods Appl. Sci. 2021, 44, 11076–11088. [Google Scholar] [CrossRef]
  11. Samoilenko, A.M.; Le Lyong, T. A method for studing boundary value problems with nonlinear boundary conditions. Ukr. Math. Zhurnal 1991, 42, 844–850. [Google Scholar] [CrossRef]
  12. Augustynovicz, A.; Kwapisz, M. On numerical-analytic method of solving of boundary value problems for functional-differential equation of neutral type. Math. Nachrichten 1990, 145, 255–269. [Google Scholar] [CrossRef]
  13. Kwapisz, M. Some remarks on integral equations arising in applications of the numerical- analytic method for the solution of boundary- value problems. Ukr. Math. Zhurnal 1992, 44, 128–132. [Google Scholar]
  14. Kwapisz, M. On modification of the integral equation of A.M. Samoilenko’s method. Math. Nachrichten 1992, 157, 125–135. [Google Scholar] [CrossRef]
  15. Samoilenko, A.M. On a convergence of polynomials and the radius of convergence of its Abel-Poisson sum. Ukr. Math. J. 2003, 55, 1119–1130. [Google Scholar] [CrossRef]
  16. Trofimchuk, E.P. Integral operators of the method of periodic successive approximations. Mat. Fiz. Nelin. Mekh. 1990, 13, 31–36. (In Russian) [Google Scholar]
  17. Mawhin, J. Topological Degree Methods in Nonlinear Boundary-Value Problems, CBMS Regional Conference Series in Mathematics; American Mathematical Society: Providence, RI, USA, 1979; Volume 40. [Google Scholar]
  18. Jankowski, T. Numerical-analytic methods for differential-algebraic systems. Acta Math. Hung. 2002, 95, 243–252. [Google Scholar] [CrossRef]
  19. Jankowski, T. Numerical-analytic methods for implicit differential equations. Math. Notes 2001, 2, 137–144. [Google Scholar] [CrossRef]
  20. Rontó, A.; Rontó, M.; Shchobak, N. On periodic solutions of systems of linear differential equations with argument deviations. J. Math. Sci. 2022, 263, 282–298. [Google Scholar] [CrossRef]
  21. Rontó, A.; Rontó, M. Successive Approximation Techniques in Non-Linear Boundary Value Problems for Ordinary Differential Equations. In Handbook of Differential Equations, Ordinary Differential Equations; Batelli, F., Feckan, M., Eds.; Elsevier B.V.: Amsterdam, The Netherlands, 2008; Volume IV, pp. 441–592. [Google Scholar]
  22. Rontó, A.; Rontó, M.; Varha, J. A new approach to non-local boundary value problems for ordinary differential systems. Appl. Math. Comput. 2015, 250, 689–700. [Google Scholar] [CrossRef]
  23. Rontó, M.; Varha, Y. Constructive existence analysis of solutions of non-linear integral boundary value problems. Miskolc Math. Notes 2014, 15, 725–742. [Google Scholar] [CrossRef]
  24. Rontó, A.; Rontó, M.; Shchobak, N. Parametrisation for boundary value problems with transcendental non-linearities using polynomial interpolation. Electron. J. Qual. Theory Differ. Equ. 2018, 59, 1–22. [Google Scholar] [CrossRef]
Figure 1. The graphs of the exact (solid line) and approximate solution (points) for m = 8 , N = 10 for the x and y components.
Figure 1. The graphs of the exact (solid line) and approximate solution (points) for m = 8 , N = 10 for the x and y components.
Axioms 13 00866 g001
Figure 2. The residual obtained as a result of substitution of the eighth approximation to exact solution for N = 10 into the given differential Equations (70) and (71) for the x and y components.
Figure 2. The residual obtained as a result of substitution of the eighth approximation to exact solution for N = 10 into the given differential Equations (70) and (71) for the x and y components.
Axioms 13 00866 g002
Figure 3. The graphs of the exact (red) and approximate second solutions (blue) for m = 9 , N = 10 for the x and y components.
Figure 3. The graphs of the exact (red) and approximate second solutions (blue) for m = 9 , N = 10 for the x and y components.
Axioms 13 00866 g003
Figure 4. The residual of the ninth approximation for the second solution for N = 10 for the x and y components.
Figure 4. The residual of the ninth approximation for the second solution for N = 10 for the x and y components.
Axioms 13 00866 g004
Table 1. Values of parameters for approximations of the exact solution.
Table 1. Values of parameters for approximations of the exact solution.
mNz γ λ
Exact 1 6 = 0.16666667 00.5
0N = 10---
0N = 60.2140434389 0.8919493467 × 10 2 0.4999204363
1N = 100.1837692491 0.1143143303 × 10 3 0.4999999869
1N = 60.1839377292 0.1197725596 × 10 3 0.4999999857
4N = 100.1623073768 0.3114629697 × 10 4 0.499999999
4N = 60.1676901732 0.9951170742 × 10 5 0.4999999999
6N = 100.1620684595 0.33438782 × 10 4 0.4999999989
6N = 60.1644399950 0.1713976081 × 10 4 0.4999999997
8N = 100.1619499438 0.3412551874 × 10 4 0.4999999988
8N = 60.1621419152 0.3262488797 × 10 4 0.4999999989
Table 2. Values of parameters for approximations of the exact solution.
Table 2. Values of parameters for approximations of the exact solution.
mNz γ λ
0N = 10−0.2229737488−0.4997757169 0.1497440580 × 10 1
0N = 6−0.2231588675−0.4997749377 0.1500038685 × 10 1
1N = 100.5140575206−0.4984984084 0.3872127112 × 10 1
1N = 60.5140119167−0.4984956534 0.3875672301 × 10 1
3N = 10−0.1050783717−0.4980623923 0.4397559982 × 10 1
3N = 6−0.1052137197−0.4980572003 0.4403436377 × 10 1
5N = 10 0.9303128843 × 10 1 −0.4982341635 0.4198473952 × 10 1
5N = 6 0.9312630660 × 10 1 −0.4982301679 0.4203212809 × 10 1
7N = 10 0.9261523108 × 10 1 −0.4982446868 0.4185967128 × 10 1
7N = 6 0.9271142850 × 10 1 −0.4982406880 0.4190724080 × 10 1
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Rontó, M. On Non-Linear Differential Systems with Mixed Boundary Conditions. Axioms 2024, 13, 866. https://doi.org/10.3390/axioms13120866

AMA Style

Rontó M. On Non-Linear Differential Systems with Mixed Boundary Conditions. Axioms. 2024; 13(12):866. https://doi.org/10.3390/axioms13120866

Chicago/Turabian Style

Rontó, Miklós. 2024. "On Non-Linear Differential Systems with Mixed Boundary Conditions" Axioms 13, no. 12: 866. https://doi.org/10.3390/axioms13120866

APA Style

Rontó, M. (2024). On Non-Linear Differential Systems with Mixed Boundary Conditions. Axioms, 13(12), 866. https://doi.org/10.3390/axioms13120866

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop