Nonlinear Higher-Order Dynamic Equation with Polynomial Growth and Mixed Boundary Conditions

Shalmali Bandyopadhyay; Svetlin G. Georgiev

REVIEW 5 major objections 5 minor 23 references

Nonlinear Higher-Order Dynamic Equation with Polynomial Growth and Mixed Boundary Conditions

T0 review · 5 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read The paper claims that a higher-order dynamic equation with mixed boundary conditions has at least one classical solution, and at least three nonnegative classical solutions under an additional ordering hypothesis.

desk verdict A genuinely new problem class in time-scale BVPs, but the central kernel is undefined and several proof steps don't hold, so the main theorems are not established as written. read the letter →

arxiv 2506.08808 v1 pith:GO5ZECH7 submitted 2025-06-10 math.CA math.DSmath.FA

Shalmali Bandyopadhyay , Svetlin G. Georgiev This is my paper

classification math.CAmath.DSmath.FA MSC 34N0539A1347H1034B15

keywords timescaleshigher-orderdynamicequationsmixedboundaryconditionspolynomialgrowthexpansivemappingscompletelycontinuousoperatorsconetheorymultiplenonnegativesolutions

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper claims existence and multiplicity for a class of nonlinear higher-order dynamic equations on a time scale, with mixed boundary conditions and polynomial growth of the nonlinear terms. The main results are Theorem 1.5, at least one classical solution under Hypothesis (A1), and Theorem 1.6, at least three nonnegative classical solutions when the extra ordering r

What carries the argument

The load-bearing object is the pair S1 and S2. S1 packages the differential expression and mixed boundary conditions into one operator-valued equation S1u=0. S2 is defined in (3.2) as S2u(t)= A/$T^{{n+1}}$ \int_0^t h_n(t,$\sigma$(s)) S1u(s) \$\Delta$ s, and Lemma 3.3 asserts that S2u identically constant implies S1u identically zero after differentiating n+1 times, so fixed points of the auxiliary equation are solutions. The fixed-point engine is Proposition 2.1 and Proposition 2.2: in the single-solution case T=eta I is expansive and I-S is compact on an equicontinuous set Y; in the three-solution case T1=(1+m)I is expansive on a cone P and S3=-|S2|-mI is completely continuous, with nested open balls U1 subset U2 subset U3 of radii r<L<R. The unstated kernel h_n must satisfy the delta-derivative identities and bounds used in Lemmas 3.2-3.3; everything else follows from norm estimates and the fixed-point theorems.

What would settle it

Find h_n. The paper never defines it, so the decisive check is to search the time-scale literature or construct it explicitly for T=$4^{{N0}}$ union {0}, n=2: one needs $\Delta$^3 integral_0^t h_2(t,$\sigma$(s)) $\varphi$(s) $\Delta$ s = $\varphi$(t) for every rd-continuous phi and |integral_0^t h_1(t,$\sigma$(s)) $\Delta$ s| <= $T^{2}$. On T=R the natural Green kernel (t-s)^n/n! would work, so the open question is whether every time scale admits such a kernel. If no kernel satisfies those identities, Lemma 3.3 fails and the claimed solutions are not established; if a kernel exists but with different bounds, Lemma 3.2's estimate ||S2u|| <= A B1 can fail and the contradiction arguments of Section 4 collapse.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that the boundary value problem (1.1) has at least one classical solution whenever the reaction term f and the boundary functions g_j satisfy the polynomial bounds in (A1), and at least three nonnegative classical solutions if, in addition, positive radii r<L<R are chosen so that the norm balls in the cone are strictly nested. The proof identifies solutions of (1.1) with zeros of the operator S1, then converts S1u=0 into the fixed-point condition S2u=0 via an (n+1)-fold integral kernel h_n; Lemma 3.3 asserts that taking n+1 delta derivatives of S2u reproduces S1u, so a fixed point of the auxiliary operator is automatically a classical solution. With that equivalence in hand, the single-solution result follows from Proposition 2.1 because the scaling part T=eta I is expansive and the correction I-S is compact on an equicontinuous set, and the three-solution result follows from Proposition 2.2 on a cone P with the completely continuous map S3=-|S2|-mI and the nested open balls U1 subset U2 subset U3. The polynomial-growth hypothesis is used only through the uniform bound ||S2u|| <= A B1 on ||u|| <= B, and the three contradiction steps in Section 4 use only that bound and the radii r<L<R. The decisive point is Lemmas 3.2-3.3: they require h_n to have the bounds and delta-derivative identities used there, but the kernel is not exhibited.

Load-bearing premise

The proof rests on the unstated existence of a kernel h_n in (3.2) that has the bounds used in Lemma 3.2 and the (n+1)-st delta-derivative reproduction identity used in Lemma 3.3; if no such kernel exists on the given time scale, S2u=0 is not equivalent to the original problem and the two main theorems do not follow.

Editorial extensions

If this is right

Every equation of the form (1.1) whose f and g_j satisfy (A1) has at least one classical solution on any time scale containing 0 and T with T>1.
Under (A2), there are at least three nonnegative classical solutions, located respectively in the norm ball of radius r, between r and L, and between L and R.
The same two-step encoding works for any order n without new estimates, provided the kernel bounds in Lemma 3.2 hold.
No monotonicity or sign condition on f is assumed, so the result admits sign-changing and non-monotone nonlinearities as long as growth is polynomial.
Because the ambient space is C^n_rd(J), the solutions are delta-differentiable to order n, so the theorem covers both differential and discrete-difference forms in one statement.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

If the absent kernel h_n is meant to be the usual Green-type kernel (in the continuous case, essentially (t-s)^n/n!), the same proof pattern would plausibly extend to systems of equations or to other boundary conditions, but the paper does not make that extension.
The nested-ball construction with three balls suggests a natural generalization to 2k+1 solutions by inserting further pairs of balls between r and R and reapplying the three-fixed-point theorem; the paper stops at three.
A reader could test the example's parameter choices on the discrete time scale 4^{N0} union {0} and look for the three claimed nonnegative solutions numerically, since the paper supplies constants but no explicit solutions.
The contradiction arguments in Section 4 use only the norm bound ||S2u|| <= A B1, not the detailed form of f and g; this leaves open the possibility of explicit radius ranges or sharper multiplicity counts, which the paper does not address.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Short version: the paper claims a new existence and triple-existence result for higher-order dynamic equations on time scales with polynomial growth and mixed boundary conditions. The combination is genuinely new, and the fixed-point setup is a reasonable template. But the central operator S2 is built on an undefined kernel h_n, and the key identity turning S2u=0 back into the original BVP is asserted without proof. As written, the theorems are not established.

What's new: the specific problem class—n-th order dynamic equation with polynomial growth in the interior and on the boundary, mixed boundary conditions, and a one-to-three solutions progression—does not appear in the papers cited. The Section 5 example is concrete and consistent with the setup.

What's good: the paper follows a clear architecture: reduce the BVP to a fixed point equation, verify the hypotheses of two known fixed point theorems, then apply them. If h_n were defined as the appropriate Taylor monomial and Lemma 3.3 proved, the reduction would likely go through. The authors also correctly identify a gap in the higher-order time-scale literature.

Where it falls down:

1. h_n is never defined. Lemma 3.2 assumes integral bounds on h_{n-k}; Lemma 3.3 claims that taking n+1 delta derivatives of (3.3) returns S1u=0. Neither the definition nor the differentiation identity is given. On an arbitrary time scale this is nontrivial. Without it, the equivalence between S2u=0 and Problem (1.1) is unverified, and both main theorems rest on that.

2. Boundary argument in Theorem 1.5: U is defined with a conditional at ∥u∥≥B/2, so ∂U includes points with ∥u∥=B/2 and u(0)=B/2. The proof asserts ∥u∥=B for all u∈∂U, which is false. The contradiction λη=1 only works for the subset with ∥u∥=B.

3. Section 4 assumptions: the paper states (I−T1)(Ω)=−(1+m)Ω; actually (I−T1)v=−mv, so the inclusion S3(U3)⊂(I−T1)(Ω) does not follow from S3u≤0. It might be repairable, but not as written.

4. Step 3: from S3(u)=(I−T1)(u−λ1u0) they get −|S2(u)|=λ1 m u0 and call it a contradiction. Both sides can be zero, so this is not a contradiction. Also u0 is introduced as an element of the dual space P* but used as a function.

5. Minor: equicontinuity of (I−S)(U) and S3(U3) is asserted rather than proved; on time scales this needs checking.

Citation pattern: Propositions 2.1 and 2.2 come from papers coauthored by Georgiev. That in itself is not a flaw—they are general published results—but it means the paper's core relies on externally quoted fixed point theorems.

Bottom line: a plausible template with real gaps. The kernel issue alone prevents acceptance. The paper needs major revision: define h_n, prove Lemma 3.3, fix the boundary sets, correct the inclusion, and rewrite Step 3. I would still send it to a referee because the problem class is legitimate, the results are new, and the errors are identifiable and possibly fixable. I would not cite it in its current form.

Referee Report

5 major / 5 minor

Summary. The paper claims to prove existence of at least one classical solution (Theorem 1.5) and at least three nonnegative classical solutions (Theorem 1.6) for a nonlinear higher-order dynamic equation on a time scale with polynomial growth and mixed boundary conditions. The proof introduces an operator S1 encoding the boundary value problem, a second operator S2 defined through an integral kernel, and applies abstract fixed-point theorems for expansive mappings combined with completely continuous operators. The abstract propositions are quoted from the literature, and the main work is to verify their hypotheses.

Significance. If the results were correct, they would contribute to the relatively sparse literature on higher-order dynamic equations with polynomial growth and mixed boundary conditions, and the proposed combination of expansive and completely continuous operators could be a useful technique. However, the manuscript contains several load-bearing gaps: the central operator S2 is defined through an unspecified kernel; the only justification for the equivalence S2u=0 with the boundary value problem is an unproved differentiation identity; compactness arguments are asserted without the necessary equicontinuity estimates; and a scalar/vector confusion appears in the verification of a boundary condition. These issues prevent the main theorems from being established as written.

major comments (5)

[§3, Eq. (3.2), Lemmas 3.2–3.3] The kernel h_n(t,σ(s)) in (3.2) is never defined. Lemma 3.2 assumes bounds of the form ∫_0^t |h_{n-k}(t,σ(s))|Δs ≤ T^{n-k+1}, and Lemma 3.3 asserts that the (n+1)-st delta derivative of S2u equals (A/T^{n+1})S1u. These are nontrivial properties that must be proved for a time-scale integral with a t-dependent kernel, particularly at right-scattered points. Without a definition of h_n and a proof of the differentiation identity, the equivalence S2u=0 ⇔ S1u=0 is unverified, and hence Theorems 1.5 and 1.6 are not established.
[Lemma 3.1] The bound |f(t,u,...)| ≤ B + ∑_{j=1}^n B^{n+1} is not justified by Hypothesis (A1). From (1.2), the j-th summand is bounded by B|Δ^{j-1}u|^{k_j} ≤ B^{1+k_j}, and there is no hypothesis ruling out k_j > n+1 or ensuring B ≤ 1. Since B1 is used in the subsequent fixed-point estimates, this gap affects the quantitative assumptions of both theorems.
[Proof of Theorem 1.5] The set U is defined using the inequality "u(0) > B/2", but u(0) is an element of R^{n+2} because u∈X=(X_1)^{n+2}; the inequality is undefined. The subsequent argument "u(0)=ληu(0) implies λη=1" relies on this undefined inequality. Interpreted as a norm inequality ∥u(0)∥>B/2 would require a specified norm and an adjustment of the definition of U. As written, the boundary contradiction is not rigorous.
[§3, p. 8 and Remark 4.1] The assertions that (I−S)(U) resides in a compact subset of Y, and that S3(U3) is relatively compact, are made after only a boundedness estimate or no supporting estimate. The Arzelà–Ascoli theorem on time scales requires equicontinuity of the family, and no equicontinuity argument is supplied. Consequently, the complete continuity hypotheses of Propositions 2.1 and 2.2 are not verified.
[§2, Propositions 2.1 and 2.2] The paper relies on fixed-point theorems quoted from [8,12] and [7,22]. Two of these sources ([8] and [12]) are listed as "Accepted" and are coauthored by one of the present authors; they are not available to the reader, and their statements are not proved or even stated precisely enough to check the hypotheses in the present setting. This is particularly problematic because conditions such as "the set {x∈∂U: x=λ(I−S)x} is empty for any λ∈(0,1/η)" are nonstandard and require careful verification.

minor comments (5)

[Lemma 3.1] In the estimate for S1ju for j∈{2,...,n}, the argument is written as |Δ^{n−j+1}u1(t) − ..., whereas the definitions in (3.1) evaluate the first term at 0 or σ^{...}(0); the indices and evaluation points should be made consistent.
[Lemma 3.2] The bound is stated for j∈{1,...,n+1} but S2 has n+2 components; the bound for j=n+2 is missing.
[Section 5] The example asserts that Hypotheses (A1) and (A2) are satisfied, but it does not verify the quantitative conditions needed for the proof, such as the smallness of η or the inequalities involving A, r, L, R, and m. Also, the equation in (5.2) is written with t∈[4,256) and Δ^2u(t/4), which does not directly match the form in (1.1).
[Proof of Theorem 1.5] The claim "S:U→X is rd-continuous and (I−S)(U) resides in a compact subset of Y" is not supported: rd-continuity of the composition with S2 is not demonstrated.
[Global] There are several minor typos, e.g., "non uniqueness in other other words" in the opening of Section 4, and the references [8] and [12] are listed as "Accepted" without journal or page numbers.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the fixed-point reformulation is a standard reduction, and the self-cited fixed-point theorems are general results independent of the target BVP.

full rationale

Circle check finds no circularity. The central chain is a standard operator reduction: (1.1) is encoded as S1u=0, then S2 is introduced as an integral of S1 with an (unstated) kernel h_n, and Lemma 3.3 inverts this integral by differentiating (n+1) times. The implication S2u=0 implies S1u=0 is not true by definition of the target objects; it requires a Green's-function identity for h_n. Because h_n is never defined and the differentiation identity is not proved, a referee cannot check the reduction, but this is an omitted-definition and omitted-proof correctness gap, not a case where the conclusion is identical to an input. The fixed-point theorems in Propositions 2.1 and 2.2, although partly self-cited from Georgiev-coauthored works, are general results about expansive and completely continuous operators in Banach spaces and cones; their stated assumptions do not include Problem (1.1) and they contain no fitted parameters, so under the independent-support rule they do not count as circularity. The cone verifications in Section 4 directly check the hypotheses of Proposition 2.2 against the BVP rather than renaming the desired conclusion as an assumption. There are serious correctness concerns (h_n undefined, the boundary of U handled incorrectly, and the example evaluating u(1) on a time scale not containing 1), but these are not circularity.

Assumptions & free parameters 3 free parameters · 4 assumptions · 2 invented entities

The proof rests on an undefined kernel h_n, on fixed point theorems quoted mostly from the authors' own accepted papers, and on informal space definitions. The extra components of the product-space fixed point are unconstrained, so the claim that a tuple solution gives a scalar solution of the BVP is an additional unstated assumption.

free parameters (3)

A = 1/(10B1)=1/40 in the example
Introduced in (3.2) to rescale the integral operator; must be small enough or combined with large m to make eta small. The example fixes A=1/40.
m = 1050 in the example
Defines T1 and S3 in (4.1); chosen large so eta=2AB1/(mr) is small. The existence proof depends on choosing m large but no explicit lower bound is given.
B = 1 in the example
Global bound on b_l, a_0j, a_kj in (A1). Appears in B1 and in the U-bound norm u < B. If B<1, the bound on f in Lemma 3.1 can fail.

assumptions (4)

standard math Propositions 2.1 and 2.2 hold as quoted.
These fixed point theorems are taken from [8,12,7,22] without proof; two sources are listed as accepted papers with S. G. Georgiev as coauthor.
ad hoc to paper The kernel h_n in (3.2) exists and satisfies the bounds and differentiation identity used in Lemmas 3.2 and 3.3.
h_n is never defined; the entire proof of Theorems depends on properties of h_n that are assumed silently.
domain assumption A fixed point in the product space X=(X1)^{n+2} corresponds to a classical solution of the scalar BVP (1.1).
The operators S1j use only the first component u1, while u2 through u_{n+2} are arbitrary; the identification of the tuple with a scalar solution is not stated or proved.
domain assumption Arzela-Ascoli applies to sets described as equi-continuous families in X.
The definitions of Y and P as sets of equi-continuous families do not form standard subspaces; the compactness claims rely on this unstated interpretation.

invented entities (2)

Kernel h_n(t,sigma(s)) in (3.2)
purpose: Builds the integral operator S2 whose zeros are meant to be solutions
Not defined in the paper; all estimates and the recovery of the equation from S2u=0 depend on its unstated properties.
Equi-continuous families space Y and P
purpose: Provides the closed or compact subset for the fixed point theorems
The text never defines what an element of Y is; equicontinuity is a property of families, not of individual functions, so the space is incoherent as written.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Nonlinear Higher-Order Dynamic Equation with Polynomial Growth and Mixed Boundary Conditions." pith.science (2026). https://pith.science/paper/GO5ZECH7

@misc{pith2026250608808,
  author       = {Pith},
  title        = {Pith review of: Nonlinear Higher-Order Dynamic Equation with Polynomial Growth and Mixed Boundary Conditions},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/GO5ZECH7}},
  note         = {Machine review of arXiv:2506.08808}
}

read the original abstract

This paper investigates the existence of solutions for a class of nonlinear higher-order dynamic equations subject to mixed boundary conditions. We consider boundary value problems in which the nonlinear reaction functions satisfy polynomial growth conditions both in the interior of the domain and on the boundary. Our analysis employs a systematic approach based on fixed-point theorems for expansive mappings combined with completely continuous operators to establish stronger existence results. Under appropriate growth conditions on the nonlinear terms, we first prove the existence of at least one classical solution, which is not guaranteed to be nonnegative. We then strengthen our hypotheses to establish the existence of at least three nonnegative solutions. The theoretical framework relies on cone theory and carefully constructed open bounded subsets within function spaces equipped with appropriate norms. Our methodology provides a unified approach to multiple solution problems for higher-order

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

23 extracted references · 23 canonical work pages

[8]

Georgiev, A

S. Georgiev, A. Kheloufi and K. Mebarki, Existence of classical solutions for a class of impulsive Hamilton-Jacobi equations, Accepted in Palestine Journal of Mathematics

work page
[12]

Khemmar, K

K. Khemmar, K. Mebarki and S. Georgiev, Existence of solutions for a class of boundary value problems for weighted p(t)-Laplacian impulsive systems, Accepted in Filomat

work page
[1]

M. A. Ahmed, A common fixed point theorem for expansive mappings in 2-metric spaces and its application, Chaos, Solitons & Fractals42(2010), 2914–2920

work page 2010
[2]

D. R. Anderson, Existence of solutions for nonlinear multi-point problems on time scales, Dynamic Systems and Applications15(2006), 21–33

work page 2006
[3]

D. R. Anderson and R. Avery, Existence of positive solutions for boundary value problems of second- order functional dynamic equations, Advances in Continuous and Discrete Models (2009), Article ID 829735

work page 2009
[4]

D. R. Anderson, Multiple positive solutions for a three-point boundary value problem, Mathematical and Computer Modelling27(1998), 49–57

work page 1998
[5]

Bohner and A

M. Bohner and A. Peterson, Dynamic equations on time scale: An Introduction with Applications, Birkh¨ auser, Boston, 2001

work page 2001
[6]

Carabello, J

T. Carabello, J. Real, and T. Taniguchi, The exponential stability of neutral stochastic delay partial differential equations, Discrete Contin. Dyn. Syst. 18, No. 2-3, 295-313 (2007). 10

work page 2007

Show all 23 references

[7]

Djebali and K

S. Djebali and K. Mebarki, Fixed point index theory for perturbation of expansive mappings by k-set contractions, Topol. Methods Nonlinear Anal., 54(2A) (2019), 613-640

2019
[9]

Henderson, Boundary value problems for systems of differential equations, boundary conditions and applications, Nonlinear Analysis7(1983), 411–417

J. Henderson, Boundary value problems for systems of differential equations, boundary conditions and applications, Nonlinear Analysis7(1983), 411–417

1983
[10]

Henderson, Solution matching for boundary value problems for linear equations, International Journal of Mathematics and Mathematical Sciences12(1989), 777–781

J. Henderson, Solution matching for boundary value problems for linear equations, International Journal of Mathematics and Mathematical Sciences12(1989), 777–781

1989
[11]

Hilger, Analysis on measure chains—a unified approach to continuous and discrete calculus, Results in Mathematics18(1990), 18–56

S. Hilger, Analysis on measure chains—a unified approach to continuous and discrete calculus, Results in Mathematics18(1990), 18–56

1990
[13]

M. A. Krasnoselskii, Two remarks on the method of successive approximations, Uspekhi Matematich- eskikh Nauk10(1955), 123–127

1955
[14]

J. W. Lyons and G. Prasad, Estimating fixed points of non-expansive mappings with an application, AIMS Mathematics6(2021), 9590–9601

2021
[15]

Ma and B

R. Ma and B. Thompson, Triple solutions to boundary value problems on time scales, Applied Mathe- matics Letters13(2000), 7–11

2000
[16]

Prasad and D

G. Prasad and D. Khantwal, Fixed Point Theorems in Relational Metric Spaces with an Application to Boundary Value Problems, Journal of Partial Differential Equations34(2021), 83–93

2021
[17]

B. E. Rhoades, Comments on two fixed point iteration methods, Journal of Mathematical Analysis and Applications56(1976), 741–750

1976
[18]

Schauder, Der Fixpunktsatz in Funktionalr¨ aumen, Studia Mathematica2(1930), 171–180

J. Schauder, Der Fixpunktsatz in Funktionalr¨ aumen, Studia Mathematica2(1930), 171–180

1930
[19]

Shahi, J

P. Shahi, J. Kaur, and S. S. Bhatia, Common Fixed Points of Expansive Mappings in Generalized Metric Spaces, Journal of Applied Research and Technology12(2014), 1168–1174

2014
[20]

Taniguchi, Common fixed point theorems on expansion type mappings on complete metric spaces, Mathematica Japonica34(1989), 139–142

T. Taniguchi, Common fixed point theorems on expansion type mappings on complete metric spaces, Mathematica Japonica34(1989), 139–142

1989
[21]

S. Z. Wang, B. Y. Li, Z. M. Gao, and K. Iseki, Some fixed point theorems on expansion mappings, Mathematica Japonica29(1984), 631–636

1984
[22]

Zahar, S

S. Zahar, S. G. Georgiev and K. Mebarki, Multiple solutions for a class of BVPs for second order ODEs via an extension of Leray-Schauder boundary condition, Nonlinear Studies 30, No. 1, 113-125 (2023)

2023
[23]

Zhou and R

W. Zhou and R. Ma, Existence and iteration of positive solutions for a generalized right-focal boundary value problem with p-Laplacian operator, Journal of Mathematical Analysis and Applications324 (2006), 409–424. 11

2006

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [8]

Georgiev, A

S. Georgiev, A. Kheloufi and K. Mebarki, Existence of classical solutions for a class of impulsive Hamilton-Jacobi equations, Accepted in Palestine Journal of Mathematics

work page

[2] [12]

Khemmar, K

K. Khemmar, K. Mebarki and S. Georgiev, Existence of solutions for a class of boundary value problems for weighted p(t)-Laplacian impulsive systems, Accepted in Filomat

work page

[3] [1]

M. A. Ahmed, A common fixed point theorem for expansive mappings in 2-metric spaces and its application, Chaos, Solitons & Fractals42(2010), 2914–2920

work page 2010

[4] [2]

D. R. Anderson, Existence of solutions for nonlinear multi-point problems on time scales, Dynamic Systems and Applications15(2006), 21–33

work page 2006

[5] [3]

D. R. Anderson and R. Avery, Existence of positive solutions for boundary value problems of second- order functional dynamic equations, Advances in Continuous and Discrete Models (2009), Article ID 829735

work page 2009

[6] [4]

D. R. Anderson, Multiple positive solutions for a three-point boundary value problem, Mathematical and Computer Modelling27(1998), 49–57

work page 1998

[7] [5]

Bohner and A

M. Bohner and A. Peterson, Dynamic equations on time scale: An Introduction with Applications, Birkh¨ auser, Boston, 2001

work page 2001

[8] [6]

Carabello, J

T. Carabello, J. Real, and T. Taniguchi, The exponential stability of neutral stochastic delay partial differential equations, Discrete Contin. Dyn. Syst. 18, No. 2-3, 295-313 (2007). 10

work page 2007