Hindawi Journal of Mathematics Volume 2022, Article ID 7668860, 17 pages https://doi.org/10.1155/2022/7668860 Research Article On Refinements of Multidimensional Inequalities of Hardy-Type via Superquadratic and Subquadratic Functions M. Zakarya ,1,2 Ghada AlNemer ,3 H. A. Abd El-Hamid,4 Roqia Butush,5 and H. M. Rezk 6 1 King Khalid University, College of Science, Department of Mathematics, P.O. Box 9004, 61413 Abha, Saudi Arabia Department of Mathematics, Faculty of Science, Al-Azhar University, 71524 Assiut, Egypt 3 Department of Mathematical Science, College of Science, Princess Nourah bint Abdulrahman University, P.O. Box 84428, Riyadh 11671, Saudi Arabia 4 Department of Mathematics and Computer Science, Faculty of Science, Beni-Suef University, Beni-Suef, Egypt 5 Department of Mathematics, University of Jordan, P.O. Box 11941, Amman, Jordan 6 Department of Mathematics, Faculty of Science, Al-Azhar University, Nasr City 11884, Egypt 2 Correspondence should be addressed to Ghada AlNemer; gnnemer@pnu.edu.sa Received 29 July 2022; Revised 15 September 2022; Accepted 10 October 2022; Published 24 November 2022 Academic Editor: Ding-Xuan Zhou Copyright © 2022 M. Zakarya et al. Tis is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. By utilizing the peculiarities of superquadratic and subquadratic functions, we give the extensions for multidimensional inequalities of Hardy-type with general kernel. We use some algebraic inequalities such as the Minkowski inequality, the refned Jensen inequality, and the Bernoulli inequality to prove the essential results in this paper. Te performance of the superquadratic functions is reliable and efective to obtain new dynamic inequalities on time scales. By utilizing special kernels, we also acquire numerous examples and implementations of the related inequalities. ∞ ∞ 1 s ds ds Φ f(t)dt ≤ Φ(f(s)) , s 0 s s 0 0 1. Introduction In [1], Hardy proved that if λ > 1, f ≥ 0 over interval (0, ∞) ∞ and 0 fλ (s)ds < ∞; then, λ λ ∞ 1 s ∞ λ λ f(t)dt ds ≤ f (s)ds, s 0 λ− 1 0 0 (1) where the constant (λ/(λ − 1))λ is sharp. By rewriting (1) with f1/λ rather than f and taking the limit as λ ⟶ ∞, we acquire the limiting case of Hardy’s inequality known as the inequality of Pólya-Knopp ([2]), that is, ∞ ∞ 1 s exp ln f(t)dtds ≤ e f(s)ds. s 0 0 0 (2) In [3], Kaijser et al. specifed that both (1) and (2) are special cases of Hardy-Knopp’s inequality. (3) where Φ ∈ C(I, R), I⊆R is a convex function and f: R+ ⟶ R+ is a locally integrable positive function. In [4], Kaijser et al. applied Fubini’s theorem and Jensen’s inequality to establish an invitingly popularization of (1). Particularly, it was evidenced that if ξ: (0, β) ⟶ R ≥ 0 and l: (0, β) × (0, β) ⟶ R ≥ 0, 0 < β ≤ ∞ such that s L(s) � l(s, t)dt > 0, 0 s ∈ (0, β), (4) and Φ ∈ C(I, R), I⊆R is a convex function and υ is defned by β υ(t) � t ξ(s) t l(s, t) ds < ∞, L(s) s Ten, the inequality, t ∈ (0, β). (5) 2 Journal of Mathematics ∞ ξ(s)Φ Al f(s) 0 ∞ ds ds ≤ υ(s)Φ(f(s)) , s s 0 holds for any non-negative integrable function f: (0, β) ⟶ R such that f(s)⊆I where Al f is defned by s 1 Al f(s) � l(s, t)f(t)dt, L(s) 0 s ∈ (0, β). (7) As a popularization of (6), Krulic et al. [5] have dem Σ, μ) are measure spaces onstrated that if (Ξ, Σ, μ) and (Ξ, with positive σ-fnite measures, ξ: Ξ ⟶ R ≥ 0 and l: Ξ × ⟶ R ≥ 0 are measurable functions such that l(s, ·) is a dμ Ξ L: Ξ ⟶ R is defned as -integrable function for s ∈Ξ, (q/p) Al f(s)dμ(s) ξ(s)Φ holds for any non-negative dμ-integrable function ⟶ R such that f(Ξ)⊆I f: Ξ where Al f: Ξ ⟶ R is defned by s ∈ Ξ. (11) In [6], the authors showed that if r ≥ 1, ξ: Ξ ⟶ R ≥ 0, ⟶ R ≥ 0 are measurable functions such that l(s, ·) l: Ξ × Ξ l(s, t) r− Φ ξ(s)Φr Al f(s)dμ(s) + r ξ(s) L(s) Ξ Ξ Ξ 1 s ∈ Ξ, (q/p) (q/p) ⎝ ξ(s)l(s, t) υ(t) � ⎛ L(s) Ξ ⎠ dμ(s)⎞ t ∈Ξ, < ∞, (9) where 0 < p ≤ q < ∞. If Φ is a convex function on I⊆R, then the inequality, (1/p) ≤ υ(s)Φ(f(s))dμ(s) Ξ (10) , L: Ξ ⟶ R is defned by is a dμ-integrable function for s ∈Ξ, (8) and the function υ is defned as r l(s, t) dμ(s) L(s) υ(t) � ξ(s) Ξ (1/r) < ∞, t ∈Ξ. (12) Moreover, if Φ is a non-negative superquadratic function, then Φ is a convex function and ξ: Ξ ⟶ R+ ≥ 0 such that υ(t) � ξ(s) Ξ l(s, t) Δμ(s) < ∞, l(s) t ∈ Ξ, (15) then (14) 1 ξ(t)Φ l(s, t)f(t)Δμ(t)Δμ(s) ≤ υ(t)Φ(f(t))Δμ(t), Ξ L(s) Ξ Ξ ⟶ R such holds for all non-negative Δμ-integrable f: Ξ ⊂ I. that f(Ξ) In [8], the authors improved the inequality (16) by replacing the function f(t) by an m-tuple of functions. f(t) � f1 (t), f2 (t), . . . , fm (t), (8) r Al f(s) × Φf(t) − Al f(s)dμ(s)dμ(t) ≤ υ(t)Φ(f(t))dμ(t) , Ξ (13) holds for any non-negative dμ-integrable function ⟶ R where Al f: Ξ ⟶ R is defned by (11). f: Ξ In [7], the researchers demonstrated some Hardy-type inequalities with a general kernel. Tey have determined that Σ, μ) are two time scale measure spaces if (Ξ, Σ, μ) and (Ξ, ⟶ R is such that with positive σ-fnite measures, l: Ξ × Ξ L(s) � l(s, t)Δμ(t) < ∞, Ξ s ∈ Ξ, and υ is defned by (1/q) Ξ 1 Al f(s) � l(s, t)f(t)dμ(t), L(s) Ξ L(s) � l(s, t)dμ(t) < ∞, Ξ (6) (17) (16) as such as f1 (t), f2 (t), . . . , fm (t) are Δμ-integrable on Ξ ⟶ R are non-negative follows: let ξ: Ξ ⟶ R, l: Ξ × Ξ such that l(s, ·) is a Δμ-integrable function for s ∈Ξ, L: Ξ ⟶ R and the function υ are defned by (14), (15), respectively. Ten, for a convex function Φ over a convex set I ⊂ Rm , the integral inequality, Journal of Mathematics 3 ξ(t)Φ Ξ 1 l(s, t)f(t)Δμ(t)Δμ(s) ≤ υ(t)Φ(f(t))Δμ(t), Ξ L(s) Ξ ⟶ Rm such that holds for all Δμ-integrable functions f: Ξ m f(Ξ) ⊂ U ⊂ R . In [9], the authors derived some inequalities of Hardytype by utilizing the concept of superquadratic functions. Particularly, they proved that if ξ: Ξ ⟶ R ≥ 0 and l: Ξ × (18) ⟶ R ≥ 0 are measurable functions such that l(s, ·) is a Δμ Ξ L: Ξ ⟶ R and the function υ -integrable function for s ∈Ξ, are defned by (14) and (15), respectively. Let Φ be a nonnegative superquadratic function. Ten, the inequality, l(s, t) Φf(t) − Al f(s)Δμ(s)Δμ(t) ≤ υ(t)Φ(f(t))Δμ(t), ξ(s)Φ Al f(s)Δμ(s) + ξ(s) Ξ L(s) Ξ Ξ Ξ holds for all non-negative Δμ-integrable ⟶ R, where Al f: Ξ ⟶ R is defned by f: Ξ Al f(s) � 1 l(s, t)f(t)Δμ(t), L(s) Ξ s ∈ Ξ. function (20) (19) Moreover, in [10, 11], the authors generalized (19) and ⟶ R ≥ 0 are proved that if ξ: Ξ ⟶ R ≥ 0 and l: Ξ × Ξ measurable functions such that l(s, ·) is a Δμ-integrable L: Ξ ⟶ R and the function υ are defned function for s ∈Ξ, by (14) and (15), respectively. Let Φ be a non-negative superquadratic function. Ten, the inequality l(s, t) Φf(t) − Al f(s)Δμ(s)Δμ(t) ≤ v(t)Φ(f(t))Δμ(t), ξ(s)Φ Al f(s)Δμ(s) + ξ(s) Ξ L(s) Ξ Ξ Ξ ⟶ Rm such that holds for all Δμ-integrable functions f: Ξ m f(Ξ) ⊂ U ⊂ R , where Al f: Ξ ⟶ R is defned by Al f(s) � 1 l(s, t)f(t)Δμ(t), L(s) Ξ s ∈ Ξ. (22) In [12], the authors deduced several generalizations of (19) on time scales. Tey proved that if r ≥ 1, ξ: Ξ ⟶ R ≥ 0 ⟶ R ≥ 0 are measurable functions such that and l: Ξ × Ξ l(s, t) r− Φ ξ(s)Φr Al f(s)Δμ(s) + r ξ(s) L(s) Ξ Ξ Ξ 1 L: Ξ ⟶ R is l(s, ·) is a Δμ-integrable function for s ∈Ξ, defned by (14) and the function υ be defned by r l(s, t) υ(t) � ξ(s) Δμ(s) L(s) Ξ (1/r) < ∞, t ∈Ξ. (23) Let Φ be a non-negative superquadratic function. Ten, the inequality, r Al f(s)Φf(t) − Al f(s)Δμ(s)Δμ(t) ≤ υ(t)Φ(f(t))Δμ(t) , (24) Ξ holds for all non-negative Δμ-integrable function ⟶ R where Al f: Ξ ⟶ R is defned by (20). f: Ξ Another development of Hardy-type inequality (24) has been made by Saker et al. [13] as follows: let ξ: Ξ ⟶ R and ⟶ R be non-negative functions such that l(s, ·) is l: Ξ × Ξ L: Ξ ⟶ R is defned by a Δμ-integrable function for s ∈Ξ, 0 < L(s) � l(s, t)Δμ t1 . . . Δμ tn < ∞, Ξ and υ is defned by (21) s � s1 , s2 , . . . , sn ∈ Ξ, (25) 4 Journal of Mathematics (1/r) r l(s, t) Δμ s1 . . . Δμ sn L(s) υ(t) � ξ(s) Ξ t � t1 , t2 , . . . , tm ∈Ξ, < ∞, (26) where r ≥ 1. If Φ is a non-negative superquadratic function, then the inequality, l(s, t) r− Φ ξ(s)Φr Al f(s)Δμ s1 . . . Δμ sn + r ξ(s) L(s) Ξ Ξ Ξ 1 Al f(s)Φf(t) − Al f(s) (27) r × Δμ s1 . . . Δμ sn Δμ t1 . . . Δμ tn ≤ υ(t)Φ(f(t))Δμ t1 . . . Δμ tn , Ξ holds for all non-negative Δμ-integrable ⟶ R where Al f: Ξ ⟶ R is defned by f: Ξ Al f(s) � 1 l(s, t)f(t)Δμ t1 . . . Δμ tn , L(s) Ξ function s ∈ Ξ. (28) In order to develop dynamic time scale inequalities, we moved the reader to the articles [14–22]. Motivated by the previous results, our major aim in this paper is to deduce several generalizations of general Hardytype inequalities for multivariate superquadratic functions that involve more general kernels on arbitrary time scales. Te paper is governed as follows: We remember some basic notions, defnitions, and results of multivariate superquadratic functions on time scales in Section 2. In Section 3, we prove some new refned dynamic inequalities of Hardy’s type with non-negative kernel by utilizing the peculiarities of superquadratic (or subquadratic) functions. In Section 4, we discuss several particular cases of Hardytype inequality by choosing such special kernels. Eventually, in Section 5, we give more implementations of our obtained results on particular time scales. 2. Preliminaries In this section, we will introduce some fundamental concepts and efects to integrals of time scales and for multivariate superquadratic functions which will be useful to deduce our major results. Before introducing the main results for multidimensional inequalities, it is necessary to present some further defnitions. Firstly, suppose m, n ∈ Z+ , Rm be the Euclidean space. Let s � s1 , s2 , . . . , sm ∈ Rm , t � t1 , t2 , . . . , tm ∈ Rm , f(t) � f1 (t), f2 (t), . . . , fm (t), be the function that is defned on t ∈ Rm . We utilize the following notations: s.t � s1 t1 , s2 t2 , . . . , sm tm , s s s s � 1 , 2 , . . . , m , t1 t2 tm t (29) analogously, 0 � (0, 0, . . . , 0) ∈ Rm is the null vector and 1 � (1, 1, . . . , 1) ∈ Rm . Correspondingly, for s, t ∈ Rm , s < t, we defne (a, b) � {t ∈ Rm : a < t < b} and the n-cells [a, b), (a, b] and [a, b] are defned similarly. Furthermore, the subsets Km and K+m in Rm are defned by Km � [0, ∞)m � t ∈ Rm : 0 ≤ t, K+m � [0, ∞)m � t ∈ Rm : 0 < t. (31) t t st � s11 s22 . . . stmm , (30) |t| � t1 , t2 , . . . , tm , m 〈s, t〉 � si ti , i�1 m also, for s, t ∈ R , we write s ≤ t(s < t) if component wise si ≤ ti (si < ti ), ∀1 ≤ i ≤ m, the relations ≥ , > rbin are defned In particular, m m 2 ⎝ t ⎞ ⎠ and t− t1 � ti , t2 � ⎛ i i�1 i�1 m 1 − 1 ⎝ t ⎞ ⎠ , �⎛ i (32) i�1 where m � (m1 , m2 , . . . , mm ). Correspondingly, [a, b) means the set [a1 , b1 ) × [a2 , b2 ) × · · · × [am , bm ), Δt � (Δt1 . . . Δtm ) and tp � (t1 . . . tm )p . Journal of Mathematics 5 Now, we arraign the defnition and some essential properties of superquadratic functions that are premised in [23]. Example 2 (see [23]). Examples 4, 5, and 6 By utilizing the same argument as in Example 4, the functions Φ1 , Φ2 , Φ3 : Km ⟶ R that defne m Defnition 1 (see [23]). A function Φ: Km ⟶ R is called superquadratic if for all s ∈ Km , there exists a function c(s) ∈ Rm such that Φ(t) − Φ(s) − Φ(|t − s|) ≥ 〈c(s), t − s〉, ∀t ∈ Km . Φ1 (t) � ti cosh ti − sinh ti , i�1 m (33) i�1 Lemma 1 (see [25]). Suppose Φ: [0, ∞) ⟶ R is continuously diferentiable and Φ(0) ≤ 0. If Φ is superadditive or ′ Φ′ (x)/x is nondecreasing, then Φ is superquadratic. In the next, we recall a couple of benefcial examples of a superquadratic function. ⎧ ⎪ 2 ⎪ ⎨ ti ln ti , Φ3 (t) � ⎪ i�1,i ≠ j ⎪ ⎩ 0, m (34) i�1 which is superquadratic on Km for each p ≥ 2 (as shown in [23], Example 2) and the function Φ: Km ⟶ R is defned by m (1/p) p ⎝ t ⎞ ⎠ Φ(t) � − ⎛ i , (35) i�1 if ti > 0, tj � 0, if t � 0, are superquadratic. Te following lemma shows that non-negative superquadratic functions are indeed convex functions: Lemma 2 (see [23]). Let Φ: Km ⟶ R be a superquadratic with c(t) � (c1 (t), c2 (t), . . . , cm (t)) that is defned as in Defnition 1. Ten, (i) Φ(0) ≤ 0 and ci (0) ≤ 0 ∀1 ≤ i ≤ m; (ii) If Φ(0) � 0 and ∇Φ(0) � (z1 Φ(0), z2 Φ(0), . . . , zm Φ(0)) � 0, then ci (t) � zi Φ(t) whenever zi Φ(t) exists for some index 1 ≤ i ≤ m at t ∈ Km ; where ∇Φ means that the gradient of the function Φ and zi Φ(t) denotes the partial derivative of Φ over the i-th variable; (iii) If Φ ≥ 0, then Φ is convex and Φ(0) � ∇Φ(0) � 0. Example 1 (see [23]). Te power function Φ: Km ⟶ R that is defned as Φ(t) � tp is called superquadratic if p ≥ 2 and subquadratic if 1 < p ≤ 2 (it is also readily seen that if 0 < p ≤ 1 then tp is a subquadratic function). Since the sum of superquadratic functions is also superquadratic, the function Φ: Km ⟶ R is defned as follows: Φ(t) � tpi , (36) i�1 m If − Φ is superquadratic, then Φ is subquadratic and the reverse inequality of (22) is held. Defnition 2 (see [24]). A function Φ: [0, ∞) ⟶ R is superadditive provided Φ(x + y) ≥ Φ(x) + Φ(y) for all x, y ≥ 0. If the reverse inequality holds, then Φ is said to be subadditive. m ⎝1 + t ⎞ ⎠ − t , Φ2 (t) � ln⎛ i i The following defnitions and theorems are referred from [26, 27]. Let T i , 1 ≤ i ≤ n be time scales, and Λn � T 1 × T 2 × · · · × T n � t � t1 , t2 , . . . , tn : ti ∈ T i , 1 ≤ i ≤ n, (37) which is called an n-dimensional time scale. Let E be Δ -measurable subset of Λn and f: E ⟶ R be a Δ-measurable function. Ten, the corresponding Δ-integral is called Lebesgue Δ-integral and is denoted by which is superquadratic on Km for each p ≥ 1 (see [23], Example 3). f t1 , t2 , . . . , tn Δ1 t1 . . . Δn tn , f(t)Δ1 t1 . . . Δn tn , fdμΔ or f(t)dμΔ (t), E E where μΔ is a σ-additive Lebesgue Δ-measure on Λn . Also, if f(t) � f1 (t), f2 (t), . . . , fm (t), (39) which is an m-tuple of functions in n-variables such that f1 , f2 , . . . , fm are Lebesgue Δ-integrable on E, then E fdμΔ denotes the m-tuple. fdμΔ 1 , . . . , fm dμΔ , E E (40) E (38) E i.e., Δ-integral acts on each component of f. Particularly, if T is an arbitrary time scale and [s, t) ⊂ T includes only isolated points, then t f(θ)Δθ � (σ(θ) − θ)f(θ) � f(θ)μ(θ), s θ∈[s,t) (41) θ∈[s,t) where μ is referred to as the graininess function on time scale. 6 Journal of Mathematics Lemma 3 (Minkowski’s inequality [13]). Let (Ξ, Σ, μ) and Σ, μ) be two fnite-dimensional time scale measures (Ξ, (1/λ) λ f(s, t)υ(t)dμ(t) ξ(s)dμ(s) Ξ Ξ provided that all integrals in (23) exist. If 0 < λ < 1, λ (43) fυdμ ξdμ > 0 and fυdμ > 0, Ξ Ξ Ξ then (42) is reversed. For λ < 0, if in addition to (43) and fλ ξdμ > 0, (44) Ξ then (42) is again reversed. ⎝ Φ⎛ E wfdμΔ E wdμΔ ⎠≤ ⎞ spaces. Suppose that ξ ≥ 0, v ≥ 0 and f ≥ 0 that are defned on and Ξ × Ξ, respectively. If λ ≥ 1, then Ξ, Ξ ≤ fλ (s, t)ξ(s)dμ(s)υ(t)dμ(t), Ξ Ξ In [25], (Corollary 5.1), Bibi get the following generalization of Jensen’s inequality and the converse of it for superquadratic functions. Theorem 1 (Jensen’s inequality). Let Φ ∈ C(Km , R) be a superquadratic function and w be Δ-integrable function for w ≥ 0 and E wdμΔ . Ten, for every Δ-integrable functions f such that f(E) ⊂ Km and wf, wΦ(f) are Δ-integrable, we have E wΦ(f)dμΔ − E wΦf − E wfdμΔ /E wdμΔ · 1dμΔ E wdμΔ . (45) ⟶ R be a non-negative so that l(s, ·) is (B2 )l: Ξ × Ξ and L: Ξ ⟶ R be Δμ-integrable function for s ∈Ξ defned by If Φ is subquadratic, then (45) is reversed. 3. Main Results In this section, we prove multidimensional Hardy-type inequalities with general kernels on time scales. Before proceeding with results, we introduce the following notations: Σ, μ) are two time scale delta (B1 )(Ξ, Σ, μ) and (Ξ, measure spaces with positive σ-fnite measures. 0 < L(s) � l(s, t)Δμ t1 . . . Δμ tn < ∞, Ξ (B3 )ξ: Ξ ⟶ R is Δμ-integrable and the function υ be defned by r l(s, t) Δμ s1 . . . Δμ sn L(s) where r ≥ 1. s ∈ Ξ. (46) (1/r) υ(t) � ξ(s) Ξ (42) < ∞, t ∈Ξ, (47) Theorem 2. Weassume (B1 ) − (B3 ) are satisfed. If Φ ∈ C(Km , R) is non-negative superquadratic function, then l(s, t) r− Φ ξ(s)Φr Al f(s)Δμ s1 . . . Δμ sn + r ξ(s) Ξ Ξ L(s) Ξ 1 Al f(s)Φf(t) − Al f(s) (48) r × Δμ s1 . . . Δμ sn Δμ t1 . . . Δμ tn ≤ υ(t)Φ(f(t))Δμ t1 . . . Δμ tn , Ξ ⟶ Rm such that holds for all Δμ-integrable functions f: Ξ ⊂ Km ⊂ Rm , where Al f: Ξ ⟶ R is defned by f(Ξ) Al f(s) � 1 l(s, t)f(t)Δμ t1 . . . Δμ tn , L(s) Ξ s ∈ Ξ. (49) Journal of Mathematics 7 By applying (45) on (50), we fnd If 0 < r < 1 and Φ is subquadratic, then (48) is reversed. Proof. We begin with the following identity: 1 Φ Al f(s) � Φ l(s, t)f(t)Δμ t1 . . . Δμ tn . L(s) Ξ (50) Φ Al f(s) + 1 1 l(s, t)Φf(t) − Al f(s)Δμ t1 . . . Δμ tn ≤ l(s, t)Φ(f(t))Δμ t1 . . . Δμ tn . L(s) Ξ L(s) Ξ (51) Since Φ ≥ 0 and r ≥ 1, we have Φ Al f(s) + r r 1 1 l(s, t)Φf(t) − Al f(s)Δμ t1 . . . Δμ tn ≤ l(s, t)Φ(f(t))Δμ t1 . . . Δμ tn . L(s) Ξ L(s) Ξ Furthermore, by employing the Bernoulli inequality [28], (1 + x)c ≤ 1 + cx, c cx + 1 − c ≤ x , for 0 < c ≤ 1 and x > − 1, for c ≥ 1 and x > 0. Φr Al f(s) + r Φr− (52) It follows that the L. H. S of (52) is not less than (53) 1 Al f(s) l(s, t)Φf(t) − Al f(s)Δμ t1 . . . Δμ tn , Ξ L(s) (54) that is, Φr Al f(s) + r Φr− r Al f(s) 1 l(s, t)Φf(t) − Al f(s)Δμ t1 . . . Δμ tn ≤ l(s, t)Φ(f(t))Δμ t1 . . . Δμ tn . Ξ L(s) L(s) Ξ 1 (55) By multiplying (55) with ξ(s) and integrating it over Ξ with respect to Δμ(s1 ) . . . Δμ(sn ), we get ξ(s)Φr Al f(s)Δμ s1 . . . Δμ sn Ξ + r ξ(s)Φr− 1 Al f(s) l(s, t)Φf(t) − Al f(s)Δμ t1 · · · μ tn × Δμ s1 . . . Δμ sn L(s) Ξ Ξ r 1 ≤ ξ(s) l(s, t)Φ(f(t))Δμ t1 . . . Δμ tn Δμ s1 . . . Δμ sn . L(s) Ξ Ξ Applying (42) on the R. H. S of (56), we get (56) 8 Journal of Mathematics r ξ(s) Ξ 1 l(s, t)Φ(f(t))Δμ t1 . . . Δμ tn Δμ s1 . . . Δμ sn L(s) Ξ r ⎝ Φ(f(t)) ξ(s)l(s, t) Δμ s . . . Δμ s ≤⎛ 1 n Ξ L(s) Ξ (57) r (1/r) ⎠. Δμ t1 . . . Δμ tn ⎞ Finally, substituting (57) into (56) and using the defnition (47) of the weight function υ, we have ξ(s)Φr Al f(s)Δμ1 s1 . . . Δμ1 sn Ξ l(s, t) r− Φ + r ξ(s) Ξ Ξ l(s) 1 Al f(s)Φf(t) − Al f(s) × Δμ s1 . . . Δμ sn Δμ t1 . . . Δμ tn (58) r ≤ υ(t)Φ(f(t))Δμ t1 . . . Δμ tn . Ξ Tis proves (48). Te proof of the case in which 0 < r < 1 and Φ is subquadratic is similar; the only diference is that the inequality sign in (48) is reversed. Te proof is complete. □ Remark 1. For m � 1, inequality (48) in Teorem 2 reduces to (27). Remark 2. For n � 1 and r � 1, inequality (48) in Teorem 2 coincides with (21). Corollary 1. Given that ξ and Al f(s) are as in Teorem 2 and ω ≥ 0 is measurable function. Since Φ ≥ 0 is superquadratic function, then the second term on the L. H. S of (48) is non-negative and the integral inequality r ξ(s)Φr Al f(s)Δμ s1 . . . Δμ sn ≤ υ(t)Φ(f(t))Δμ t1 . . . Δμ tn , Ξ Ξ holds. Remark 3. By taking r � 1 and n � 1, inequality (59) in Corollary 1 reduces to (18). (59) Remark 4. For m � 1 and n � 1, inequality (59) in Corollary 1 coincides with inequality (2.2) which is [29], (Corollary 2.1.2). Remark 5. If we rewrite (48) with r � λ/μ ≥ 1, 0 < μ ≤ λ < ∞ or − ∞ < μ ≤ λ < 0 and Φ ≥ 0, then we get ξ(s)Φ(λ/μ) Al f(s)Δμ s1 . . . Δμ sn Ξ λ l(s, t) (λ/μ)− Φ + ξ(s) μ Ξ Ξ L(s) 1 Al f(s)Φf(t) − Al f(s) (60) (λ/μ) × Δμ s1 . . . Δμ sn Δμ t1 . . . Δμ tn ≤ υ(t)Φ(f(t))Δμ t1 . . . Δμ tn Ξ Remark 6. For m � 1 and n � 1, inequality (60) coincides with inequality (3.13) which is [27], (Remark 3.5). . Remark 7. In Remark 5, since Φ ≥ 0, then the second term on the L. H. S. of (60) is non-negative and (60) Journal of Mathematics 9 reduces to the weighted Hardy-type inequality of the form ξ(s)Φ(λ/μ) Al f(s)Δμ s1 . . . Δμ sn ≤ υ(t)Φ(f(t))Δμ t1 . . . Δμ tn Ξ Ξ which is a refnement of general Hardy-type inequality established in [29] [Remark 2.1.4] and [5]. As a specifc case of Teorem 2 when Φ(t) � tλ for λ ≥ 2, we get the next result. (λ/μ) (61) , Corollary 2. Suppose that the assumptions of Teorem 2 are satisfed and λ ≥ 2. Ten, λ l(s, t) λr λ(r− 1) Al f(s) ξ(s) Al f(s) Δμ s1 . . . Δμ sn + r ξ(s) f(t) − Al f(s) × Δμ s1 . . . Δμ sn Δμ t1 . . . Δμ tn L(s) Ξ Ξ Ξ λ (62) r ≤ υ(t)f (t)Δμ t1 . . . Δμ tn . Ξ If 0 < r < 1 and 1 < λ ≤ 2, then (62) is reversed. Remark 8. Clearly, for m � 1, inequality (62) in Corollary 2 coincides with inequality (46) which is [13], (Corollary 2.1). In fact, the function Φ(s) � es is not superquadratic but by working with the superquadratic function Φ(s) � es − s − 1 (see Lemma 3) and replacing f(s) by ln f(s) in Teorem 2, we obtain the next multidimensional version of the Pólya–Knopp type inequality. Corollary 3. Suppose that the assumptions in Teorem 2 are satisfed and assume that Al f(s) � 1 l(s, t)ln f(t)Δμ t1 . . . Δμ tn , L(s) Ξ s ∈ Ξ, (63) then r r ξ(s) exp Al f(s) − Al f(s) − 1 Δμ s1 . . . Δμ sn + I ≤ υ(t)(f(t) − ln f(t) − 1)Δμ t1 . . . Δμ tn , Ξ Ξ (64) l(s, t) r− 1 I � r ξ(s) exp Al f(s) − Al f(s) − 1 L(s) Ξ Ξ × expln f(t) − Al f(s) − lnf(t) − Al f(s) − 1 × Δμ s1 . . . Δμ sn Δμ t1 . . . Δμ tn . (65) where If 0 < r < 1, then (64) is reversed. Remark 9. For m � 1, inequality (64) in Corollary 3 coincides with inequality (48) which is [13], (Corollary 2.2). In the next results, we also suppose the following hypothesis: (B1′) Let � [a1 , β1 )T × [a2 , β2 )T × · · · × [an , βn )T ⊂ Rn , Ξ�Ξ 0 ≤ ai < βi ≤ ∞, for every 1 ≤ i ≤ n, where T is an arbitrary time scale and let Δμ1 (s) � Δs, Δμ2 (t) � Δt. Theorem 3. Weassume (B1′) and (B2 ) are satisfed. Suppose τ: Ξ ⟶ R+ such that r ω(t) � s1 . . . sn 1 l(s, t) τ(s) Δs1 . . . Δsn L(s) Ξ σ s 1 . . . σ sn (1/r) , t ∈ Ξ, (66) 10 Journal of Mathematics where r ≥ 1. If Φ ∈ C(Km , R) is non-negative superquadratic function, then β1 βn a1 an . . . τ(s)Φr A∗l f(s) β1 βn a1 an β1 βn t1 tn Δs1 . . . Δsn σ s1 . . . σ s n + r . . . . . . τ(s) l(s, t) r− Φ L(s) 1 A∗l f(s)Φf(t) − A∗l f(s) (67) r × β1 βn Δs1 . . . Δsn Δt1 . . . Δtn ≤ . . . ω(t)Φ(f(t))Δt1 . . . Δtn , σ s1 . . . σ s n a1 an holds for all Δμ-integrable functions f: Ξ ⟶ Rm such that f(Ξ) ⊂ Km ⊂ Rm , where A∗l f: Ξ ⟶ R is defned by A∗l f(s) � β1 βn 1 . . . l(s, t)f(t)Δt1 . . . Δtn , L(s) a1 an Remark 10. For r � 1, Teorem 3 coincides with [10], (Corollary 2.11). s ∈ Ξ. Remark 11. In Teorem 3, if we replace ω(t) by ω(t)/(s1 . . . sn ) and put m � 1, then we get the result given in [13], (Corollary 2.3). (68) If Φ is subquadratic function and 0 < r < 1, then (48) is reversed. 4. Inequalities with Special Kernels In this section, we get some consequential inequalities of Hardy-type by selecting special kernels. Proof. We get the result from Teorem 2 by taking ξ(s) � τ(s) . σ s 1 . . . σ sn (69) □ β1 βn t1 tn ω2 (t) � . . . ξ(s) Theorem . Suppose (B1′) and 1 r ni�1 σ si − ai Δs1 . . . Δsn )(1/r) < ∞, (70) such that ξ: Ξ ⟶ R+ is Δμ-integrable function and r ≥ 1. If Φ ∈ C(Km , R) is non-negative superquadratic function, then β1 βn a1 an . . . ξ(s)Φr Al∗ ∗ f(s)Δs1 . . . Δsn β1 βn a1 an β1 βn t1 tn + r ... ... ni�1 ξ(s) Φr− σ si − a i β1 βn a1 an 1 Al∗ ∗ f(s)Φf(t) r ≤ . . . ω2 (t)Φ(f(t))Δt1 . . . Δtn , holds for all Δμ-integrable functions f: Ξ ⟶ Rm such that f(Ξ) ⊂ Km ⊂ Rm , where Al∗ ∗ f: Ξ ⟶ R is defned by − Al∗ ∗ f(s) (71) × Δs1 . . . Δsn Δt1 . . . Δtn Journal of Mathematics 11 Al∗ ∗ f(s) � σ (s1 ) σ (sn ) 1 ... f(t)Δt1 . . . Δtn , σ si − a i a 1 an ni�1 s ∈ Ξ. (72) If Φ is subquadratic function and 0 < r < 1, then (71) is reversed. Remark 12. For r � 1, Teorem 4 coincides with [10], (Corollary 3.1). Proof. We get the result from Teorem 2 by taking We have the following in this case: Remark 13. For m � 1, Teorem 4 coincides with [13], (Corollary 2.7). If we let ai � 0, (i � 1, . . . , n) and ξ(s) � 1/(s1 . . . sn ) in Teorem 4, we have the next result. L(s) � σ (s 1 ) a1 σ (s n ) ... an n Δt1 . . . Δtn � σ si − ai , (73) i�1 Al � Al∗ ∗ , ω � ω2 . □ β1 βn t1 tn ω3 (t) � . . . Corollary . Assuming (B1′), we defne 1 r Δs1 . . . Δsn s1 . . . sn ni�1 σ si (1/r) , t ∈ Ξ, (74) such that ξ: Ξ ⟶ R+ is Δμ-integrable function and r ≥ 1. If Φ ∈ C(Km , R) is non-negative superquadratic function, then β1 βn 0 0 . . . Φr A′lf(s) Δs1 . . . Δsn ni�1 si β1 βn β1 βn 0 0 t1 tn + r . . . . . . Φr− 1 A′lf(s)Φf(t) − A′lf(s) (75) r × β1 βn Δs1 . . . Δsn Δt . . . Δtn ≤ . . . ω3 (t)Φ(f(t))Δt1 . . . Δtn , ni�1 si σ si 1 0 0 holds for all Δμ-integrable functions f: Ξ ⟶ Rm such that f(Ξ) ⊂ Km ⊂ Rm , where A′lf: Ξ ⟶ R is defned by A′lf(s) � σ (s1 ) σ (sn ) 1 ... f(t)Δt1 . . . Δtn , σ si 0 0 If Φ is subquadratic function and 0 < r < 1, then (76) is reversed. s ∈ Ξ. ni�1 Remark 14. If we take βi � ∞(i � 1, . . . , n), then inequalities (75) and (76) reduces to (76) ∞ (1/r) ∞ ω3 (t) � . . . t1 ∞ ∞ 0 0 1 r Δs1 . . . Δsn s1 . . . sn ni�1 σ si tn . . . Φr A′lf(s) ∞ 0 t ∈ Ξ, Δs1 . . . Δsn ni�1 si ∞ ∞ ∞ 0 t1 tn + r . . . . . . Φr− ∞ , 1 (77) Δs . . . Δsn A′lf(s)Φf(t) − AA′lf(s) × n1 Δt . . . Δtn i�1 si σ si 1 ∞ r ≤ . . . ω3 (t)Φ(f(t))Δt1 . . . Δtn . 0 0 Remark 15. Clearly, for r � 1, inequalities (75) and (76), respectively, reduces to β1 βn t1 tn ω3 (t) � . . . n 1 1 1 Δs . . . Δs � − , n 1 n ti βi i�1 si σ si i�1 (78) 12 Journal of Mathematics and β1 βn 0 0 ... 1 Φ A′lf(s)Δs1 . . . Δsn ni�1 si β1 βn β1 βn 0 0 t1 tn β1 βn n 0 0 i�1 + ... ... 1 1 1 Φf(t) − A′lf(s)Δs1 . . . Δsn Δt1 . . . Δtn − Φ(f(t))Δt1 . . . Δtn . ni�1 si σ si ti βi (79) ≤ ... Furthermost, if we take βi � ∞(i � 1, . . . , n), then (80) becomes ∞ ∞ ... 0 ∞ 1 n 0 i�1 si ∞ ∞ ∞ Φ A′lf(s)Δs1 . . . Δsn + . . . . . . 0 0 t1 tn ∞ ∞ n 1 Φf(t) − A′lf(s) × Δs1 . . . Δsn Δt1 . . . Δtn ≤ . . . ti , n i�1 si σ si 0 0 i�1 Theorem 5. Assume (B1′) with βi � ∞(i � 1, . . . , n). If Φ ∈ C(Km , R) is non-negative superquadratic function, then which is [10], [Remark 3.2]. ∞ ∞ n a1 an i�1 ∞ ∞ s1 sn ⎝ s . . . . . . Φr ⎛ i ∞ ∞ σ (t 1 ) + r ... a1 (80) an σ (t n ) ... a1 an f(t) ⎠ Δs1 .n. . Δsn Δt . . . Δtn ⎞ ni�1 ti σ ti 1 i�1 si n ∞ ∞ s1 sn ⎝ s . . . Φr− 1 ⎛ i i�1 f(t) ⎠ Δt . . . Δtn ⎞ ni�1 ti σ ti 1 n ∞ ∞ f(t) ⎝ ⎞ × Φ⎛f(t) − si . . . Δt1 . . . Δtn ⎠ n s1 sn i�1 ti σ ti i�1 (81) r × Δs1 . . . Δsn ∞ ∞ Δt1 . . . Δtn ≤ . . . ω(t)Φ(f(t))Δt1 . . . Δtn , n i�1 ti σ ti a1 an holds for all Δμ-integrable functions f: Ξ ⟶ Rm such that f(Ξ) ⊂ Km ⊂ Rm , where ∞ ∞ a1 n an i�1 si ⎝ . . . ω(t) � ⎛ 1 ⎝ ⎛ 1/ni�1 ti σ ti r 1/ni�1 si Δs1 . . . Δsn n If Φ is subquadratic function and 0 < r < 1, then (82) is reversed. Proof. We get the result from Teorem 2 by taking l(s, t) � i�1 (1/r) , t ∈ Ξ. (82) 1 , if si ≥ t for all i ∈ {1, . . . , n}, 0, otherwise. t i σ ti (83) and ξ(s) � 1/(s1 . . . sn ). □ Journal of Mathematics 13 Corollary 5. Clearly, for r � 1, inequality (82) reduces to ∞ ∞ a1 an n ∞ ∞ s1 sn ⎝ s . . . . . . Φ⎛ i i�1 ∞ σ (t 1 ) ∞ + ... a1 an × Δs1 . . . Δsn f(t) ⎠ Δs1 .n. . Δsn Δt . . . Δtn ⎞ Πi�1 si ni�1 ti σ ti 1 σ (t n ) ... a1 an n ∞ ∞ f(t) ⎝ ⎠ ⎛ Δt1 . . . Δtn ⎞ Φ f(t) − si . . . n s1 sn i�1 ti σ ti i�1 (84) ∞ ∞ n Δt1 . . . Δtn ai Δt . . . Δtn ≤ . . . 1 − , Φ(f(t)) 1 n n σ t i�1 ti σ ti i�1 ti a1 an i�1 i Theorem 6. Assume T be an isolated time scale, ξ: Ξ ⟶ R+ � [a, ∞)T , a ≥ 0. If λ ≥ 1, is Δμ-integrable function and Ξ � Ξ then which is [10], (Corollary 3.3). 5. Inequalities with Particular Time Scales In this section, we get some consequential inequalities by selecting some specifc time scales. λ (r/λ) λ ξ(s) A″l f1 (s) + A″l f2 (s) μ(s) + r s∈[a,∞)T t∈[a,∞)T s∈[t,∞)T (r− 1/λ) ξ(s) λ λ A″l f1 (s) + A″l f2 (s) σ(s) − a λ (1/λ) λ ′′ μ(s)μ(t) f1 (t) − A″l f1 (s) + f2 (t) − Al f2 (s) (85) r ω4 (t)fλ1 (t) ⎝ ≥⎛ + (1/λ) ⎠ fλ2 (t) μ(t)⎞ , t ∈ [a,∞)T ⟶ R+ , where holds for all Δμ-integrable functions f1 , f2 : Ξ 1 A″l fi (s) � σ(s) − a fi (t)μ(t), i � 1, 2, (86) Proof. We get the result from Teorem 4 by taking (1/λ) m ⎝ sλ ⎞ ⎠ Φ(s) � − ⎛ i t∈[a,σ(s))T (88) , i�1 and □ with λ ≥ 1, n � 1 and m � 2. ⎝ ω4 (t) � ⎛ s ∈ [t,∞)T ′ λ ξ(s) ⎠ μ(s)⎞ (σ(s) − a)r ′ λ ξ(s)A′l f1 (s) + A′l f2 (s) (1/r) . (87) Remark 16. If we take r � 1 in Teorem 6, then we get (1/λ) μ(s) s∈[a,∞)T + t∈[a,∞)T s∈[t,∞)T λ λ (1/λ) ξ(s) μ(s)μ(t) ≥ f1 (t) − A″l f1 (s) + f2 (t) − A″l f2 (s) σ(s) − a ω5 (t)fλ1 (t) + fλ2 (t) (1/λ) μ(t), t∈[a,∞)T (89) where (A″l fi )(s), i � 1, 2 is defned by (87) and 14 Journal of Mathematics ω5 (t) � s∈[t,∞)T ξ(s) μ(s), σ(s) − a (90) Theorem 7. Assuming T be an isolated time scale, � [a, ∞)T , ξ: Ξ ⟶ R+ is Δμ-integrable function and Ξ � Ξ a ≥ 0. Ten, which is [10], (Corollary 4.1). ∧ ∧ ∧ ∧ ∧ ∧ ξ(s)μ(s)Al f1 (s)coshAl f1 (s) − sinhAl f1 (s)+ + Al f2 (s)coshAl f2 (s) − sinhAl f2 (s)r + r s∈[a,∞)T r− ∧ ∧ ∧ ∧ ∧ ∧ ξ(s) Al f1 (s)coshAl f1 (s) − sinhAl f1 (s) + Al f2 (s)coshAl f2 (s) − sinhAl f2 (s) σ(s) − a t∈[a,∞)T s∈[t,∞)T 1 ∧ ∧ ∧ × f1 (t) − Al f1 (s)coshf1 (t) − Al f1 (s) − sinhf1 (t) − Al f1 (s) 1 ∧ ∧ ∧ + f2 (t) − Al f2 (s)coshf2 (t) − Al f2 (s) − sinhf2 (t) − Al f2 (s)μ(s)μ(t) r ⎝ ≤⎛ ⎠, ω6 (t) f1 (t)cosh f1 (t) − sinh f2 (t) + f2 (t)cosh f2 (t) − sinh f2 (t)μ(t)⎞ t ∈ [a,∞)T (91) ⟶ R+ , where holds for all Δμ-integrable functions f1 , f2 : Ξ l fi (s) � A 1 σ(s) − a Proof. We get the result from Teorem 4 by taking m fi (t)μ(t), i � 1, 2, (92) t∈[a,σ(s))T Φ(s) � si coshsi − sinhsi , (94) i�1 and □ with n � 1 and m � 2. (1/r) ⎝ ω6 (t) � ⎛ s ∈ [t,∞)T ξ(s) ⎠ μ(s)⎞ (σ(s) − a)r . (93) Remark 17. If we take r � 1 in Teorem 7, then we get l f1 (s) − sinhA l f1 (s) + A l f2 (s)coshA l f2 (s) − sinhA l f2 (s) l f1 (s)coshA ξ(s)μ(s)A s∈[a,∞)T + t∈[a,∞)T s∈[t,∞)T ξ(s) l f1 (s)coshf1 (t) − A l f1 (s) − sinhf1 (t) − A l f1 (s) f1 (t) − A σ(s) − a (95) l f2 (s)coshf2 (t) − A l f2 (s) − sinhf2 (t) − A l f2 (s)μ(s)μ(t) + f2 (t) − A ≤ ω7 (t) f1 (t)cosh f1 (t) − sinh f1 (t) + f2 (t)cosh f2 (t) − sinh f2 (t)μ(t), t∈[a,∞)T l fi )(s), i � 1, 2 is defned by (93) and where (A ω7 (t) � s∈[t, ∞)T ξ(s) μ(s), σ(s) − a which is [10], (Corollary 4.3). (96) Theorem 8. Assuming T be an isolated time scale, � [a, ∞)T , ξ: Ξ ⟶ R+ is Δμ-integrable function and Ξ � Ξ a ≥ 0. Ten, Journal of Mathematics 15 2 r 2 ξ(s) Al f1 (s)ln Al f1 (s) + Al f2 (s)ln Al f2 (s) μ(s) s∈[a,∞)T +r t∈[a,∞)T s∈[t,∞)T r− ξ(s) 2 2 Al f1 (s)ln Al f1 (s) + Al f2 (s)ln Al f2 (s) σ(s) − a 1 2 2 × f1 (t) − Al f1 (s) lnf1 (t) − Al f1 (s) + f2 (t) − Al f2 (s) lnf2 (t) − Al f2 (s)μ(s)μ(t) (97) r ⎝ ≤⎛ ⎠, ω8 (t)f21 (t)ln f1 (t) + f22 (t)ln f2 (t)μ(t)⎞ t ∈ [a,∞)T ⟶ R+ , where holds for all Δμ-integrable functions f1 , f2 : Ξ Al fi (s) � 1 σ(s) − a Proof. We get the result from Teorem 4 by taking m f(t)μ(t)i , i � 1, 2, Φ(s) � s2i ln si , (98) t∈[a,σ(s))T (100) i�1 and □ with the assumption 0 ln 0 � 0. (1/r) ⎝ ω9 (t) � ⎛ s ∈ [t,∞)T ξ(s) ⎠ μ(s)⎞ (σ(s) − a)r . (99) 2 Remark 18. If we take r � 1 in Teorem 8, then we get 2 ξ(s) Al f1 (s)ln Al f1 (s) + Al f2 (s)ln Al f2 (s)μ(s) s∈[a,∞)T f1 (t) − Al f1 (s)2 lnf1 (t) − Al f1 (s) + ≤ ξ(s) ⎜ ⎟ ⎛ ⎞ ⎜ ⎟ ⎜ ⎟ ⎝ ⎠μ(s)μ(t) 2 σ(s) − a t∈[a,∞)T s∈[t,∞)T +f2 (t) − Al f2 (s) lnf2 (t) − Al f2 (s) (101) ω9 (t)f21 (t)ln f1 (t) + f22 (t)ln f2 (t)μ(t), t∈[a,∞)T where (Al fi )(s), i � 1, 2 is defned by (99) and ω9 (t) � s∈[t,∞)T ξ(s) μ(s), σ(s) − a (102) Theorem 9. Assuming T be an isolated time scale, � [a, ∞)T , ξ: Ξ ⟶ R+ is Δμ-integrable function and Ξ � Ξ a ≥ 0. Ten, which is [6], (Corollary 4.4). r ξ(s)ln1 + A€l f1 (s) + A€l f2 (s) − A€l f1 (s) − A€l f2 (s) μ(s) s∈[a,∞)T +r t∈[a,∞)T s∈[t,∞)T ξ(s) r− ln1 + A€l f1 (s) + A€l f2 (s) − A€l f1 (s) − A€l f2 (s) σ(s) − a 1 × ln1 + f1 (t) − A€l f1 (s) + f2 (t) − A€l f2 (s) − f1 (t) − A€l f1 (s) − f2 (t) − A€l f2 (s)μ(s)μ(t) r ⎝ ≤⎛ t ∈ [a,∞)T ⎠, ω10 (t)ln 1 + f1 (t) + f2 (t) − f1 (t) − f2 (t)μ(t)⎞ (103) 16 Journal of Mathematics holds for all Δμ-integrable functions f1 , f2 : Ξ2 ⟶ R+ , where 1 A€l fi (s) � σ(s) − a f(t)i μ(t), i � 1, 2, m m ⎝1 + s ⎞ ⎠ − s . Φ(s) � ln⎛ i i (104) i�1 (106) i�1 □ t∈[a,σ(s))T and ⎝ ω10 (t) � ⎛ Proof. We get the result from Teorem 4 by taking s ∈ [t,∞)T ξ(s) ⎠ μ(s)⎞ (σ(s) − a)r Remark 19. If we take r � 1 in Teorem 9, then we get (1/r) . (105) ξ(s)ln1 + A€l f1 (s) + A€l f2 (s) − A€l f1 (s) − A€l f2 (s)μ(s) s∈[a,∞)T + t∈[a,∞)T s∈[t,∞)T ≤ ξ(s) ln(1 + f1 (t) − A€l f1 (s) + f2 (t) − A€l f2 (s) − f1 (t) − A€l f1 (s) − f2 (t) − A€l f2 (s)μ(s)μ(t) σ(s) − a ω11 (t)ln 1 + f1 (t) + f2 (t) − f1 (t) − f2 (t)μ(t), t∈[a,∞)T (107) where (A€l fi )(s), i � 1, 2 is defned by (105) and ω11 (t) � s∈[t,∞)T ξ(s) μ(s), σ(s) − a Authors’ Contributions (108) which is [10], (Corollary 4.5). H. M. Rezk, H. A. Abd El-Hamid, R. Butush, and G. AlNemer performed software analysis and wrote the original draft. H. M. Rezk and M. Zakarya reviewed and edited the manuscript. All the authors have read and agreed to the published version of the manuscript. 6. Conclusion and Future Work Tis research article is dedicated for some general dynamic inequalities of Hardy’s type and their converses on time scales. Tese inequalities are considered rather in general terms and contain a number of special integral inequalities. In particular, our fndings can be seen as refnements of some recent results closely linked to classical Hardy and Pólya–Knopp inequalities on time scale. We use some algebraic inequalities such as the Minkowski inequality, the refned Jensen inequality, and the Bernoulli inequality on time scales to prove the essential results in this paper. Our computed outcomes can be very useful as a starting point to get some continuous inequalities as special cases. In the future, such inequalities can be introduced by using fractional integrals and fractional derivatives of the Riemann− Liouville type on time scales. It will also be very enjoyable to introduce such inequalities in quantum calculus. Data Availability No data were used to support this study. Conflicts of Interest Te authors declare that they have no conficts of interest. Acknowledgments Princess Nourah bint Abdulrahman University Researchers Supporting Project number (PNURSP2022R45), Princess Nourah bint Abdulrahman University, Riyadh, Saudi Arabia. References [1] G. H. Hardy, “Notes on some points in the integral calculus, LX an inequality between integrals,” Messenger of Math, vol. 54, pp. 150–156, 1925. [2] K. Knopp, “Über Reihen mit positiven Gliedern,” Journal of the London Mathematical Society, vol. s1-3, pp. 205–211, 1928. [3] S. Kaijser, L. E. Persson, and A. Öberg, “On carleman and knopp’s inequalities,” Journal of Approximation Teory, vol. 117, pp. 140–151, 2002. [4] S. Kaijser, L. Nikolova, L. E. Persson, and A. Wedestig, “Hardy type inequalities via convexity,” Mathematical Inequalities and Applications, vol. 8, no. 3, pp. 403–417, 2005. [5] K. K. H. lreich, J. E. Pečarić, and L. E. Persson, “Some new Hardy type inequalities with general kernels,” Mathematical Inequalities and Applications, vol. 12, no. 3, pp. 473–485, 2009. [6] A. Čižmešija, J. E. Pečarić, and L. E. Persson, “On strengthened Hardy and Pólya-Knopp’s inequalities,” Journal of Approximation Teory, vol. 125, no. 1, pp. 74–84, 2003. Journal of Mathematics [7] M. J. Bohner, A. Nosheen, J. E. Pečarić, and A. Younus, “Some dynamic Hardy-type inequalities with general kernel,” Journal of Mathematical Inequalities, vol. 8, no. 1, pp. 185–199, 2014. [8] T. Donchev, A. Nosheen, and J. E. Pečarić, “Hardy-type inequalities on time scale via convexity in several variables,” ISRN Mathematical Analysis, vol. 2013, Article ID 903196, 9 pages, 2013. [9] J. Adedayo Oguntuase and L. E. Persson, “Time scales Hardytype inequalities via superquadracity,” Annals of Functional Analysis, vol. 5, no. 2, pp. 61–73, 2014. [10] R. Bibi and W. Ahmad, “Hardy type inequalities for several variables and their converses for time scales integrals via superquadratic functions,” Analysis Mathematica, vol. 45, no. 2, pp. 249–265, 2019. [11] O. O. Fabelurin and J. A. Oguntuase, “Multivariat Hardy-type inequalities on time scales via superquadraticity,” Proceedings of A. Razmadze Mathematical Institute, vol. 167, pp. 29–42, 2015. [12] S. H. Saker, H. M. Rezk, and M. Krnić, “More accurate dynamic Hardy-type inequalities obtained via superquadraticity,” Revista de la Real Academia de Ciencias Exactas, Fı́sicas y Naturales Serie A. Matemáticas, vol. 113, no. 3, pp. 2691–2713, 2019. [13] S. H. Saker, H. M. Rezk, I. Abohela, and D. Baleanu, “Refnement multidimensional dynamic inequalities with general kernels and measures,” Journal of Inequalities and Applications, vol. 306, pp. 1–16, 2019. [14] S. Abramovich, G. Jameson, and G. Sinnamon, “Refning of Jensen’s inequality,” Bull. Math. Soc. Sci. Math. Roumanie (N.S.), vol. 47, no. 95, pp. 3–14, 2004. [15] H. A. A. El-Hamid, H. M. Rezk, A. M. Ahmed, G. AlNemer, M. Zakarya, and H. A. El Saify, “Dynamic inequalities inquotients with general kernels and measures,” Journal of Function Spaces, vol. 2020, Article ID 5417084, 12 pages, 2020. [16] J. Barić, R. Bibi, M. Bohner, and J. Pečarić, “Time scales integral inequalities for superquadratic functions,” Journal of the Korean Mathematical Society, vol. 50, no. 3, pp. 465–477, 2013. [17] H. Feng, S. Hou, L. Y. Wei, and D. X. Zhou, “CNN models for readability of Chinese texts,” Mathematical Foundations of Computing, vol. 5, no. 4, pp. 351–362, 2022. [18] H. Karsli, “On multidimensional Urysohn type generalized sampling operators,” Mathematical Foundations of Computing, vol. 4, no. 4, pp. 271–280, 2021. [19] R. Liu and R. Xu, “Hermite-Hadamard type inequalities for harmonical (h1, h2)-convex interval-valued functions,” Mathematical Foundations of Computing, vol. 4, no. 2, pp. 89–103, 2021. [20] D. O’Regan, H. M. Rezk, and S. H. Saker, “Some dynamic inequalities involving Hilbert and Hardy-Hilbert operators with kernels,” Results in Mathematics, vol. 73, no. 4, pp. 146–222, 2018. [21] H. M. Rezk, H. A. Abd El-Hamid, A. M. Ahmed, G. AlNemer, and M. Zakarya, “Inequalities of Hardy type via superquadratic functions with general kernels and measures for several variables on time scales,” Journal of Function Spaces, vol. 2020, Article ID 6427378, 15 pages, 2020. [22] D. X. Zhou, “Deep distributed convolutional neural networks: universality,” Analysis and Applications, vol. 16, no. 6, pp. 895–919, 2018. [23] S. Abramovich, S. Banić, and M. Matić, “Superquadratic functions in several variables,” Journal of Mathematical Analysis and Applications, vol. 327, no. 2, pp. 1444–1460, 2007. 17 [24] S. Abramovich, K. Krulic, J. Pečarić, and L.-E. Persson, “Some new refned Hardy type inequalities with general kernels and measures,” Aequationes Mathematicae, vol. 79, no. 1-2, pp. 157–172, 2010. [25] R. Bibi, “Jessen type inequalities for several variables via superquadratic functions,” Banach Journal of Mathematical Analysis, vol. 10, no. 1, pp. 120–132, 2016. [26] M. Bohner and A. Peterson, Dynamic Equations on Time Scales: An Introduction with Applications, Birkhäuser, Boston, MA, USA, 2001. [27] M. Bohner and A. Peterson, Eds., Advances in Dynamic Equations on Time Scales, Birkhäuser, Boston, MA, USA, 2003. [28] D. S. Mitrinović, J. E. Pečarić, and A. M. Fink, Classical and New Inequalities in Analysis, Kluwer Academic Publishers, Amsterdam, Netherlands, 1993. [29] K. Rauf and O. A. Sanusi, “Weighted hard-type inequalities on time scales via superquadraticity,” Journal of Mathematical Analysis and Applications, vol. 8, pp. 107–122, 2017.