An inequality for Jacobi polynomials: a complement to Finite Increment Theorem

Geno Nikolov · 2026

Read An inequality for Jacobi polynomials: a complement to Finite Increment Theorem by Geno Nikolov (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2606-25468.txt.

Introduction

Jacobi polynomials are a classical family weighted differently at the two ends of an interval, and their shape obeys tight inequalities. Geno Nikolov proves a new one in 2026 that reads like a sharpened mean value statement: the drop of the polynomial from its endpoint value bounds its slope times the remaining distance. The result holds on the right half of the interval for two ranges of the weight parameters, including an endpoint case tied to Chebyshev polynomials.

The inequality complements an earlier finite increment theorem for the same family rather than restating it. Nikolov's route builds on nonnegative expansions developed in prior work on trigonometric inequalities by Robertson and by Askey and Gasper. Special cases recover statements about Chebyshev polynomials of both kinds, which sit inside the Jacobi family at particular parameter choices.

Catalog record

CREATORGeno Nikolov
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE14 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2606.25468

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

An inequality for Jacobi polynomials: a complement to Finite Increment Theorem Geno Nikolov Abstract . Let P = P n ( α , β ) P=P_{n}^{(\alpha,\beta)} be the n n -th degree Jacobi polynomial, which is orthogonal in [ − 1 , 1 ] [-1,1] with respect to the weight function ( 1 − x ) α ​ ( 1 + x ) β (1-x)^{\alpha}(1+x)^{\beta} , α , β > − 1 \alpha,\beta>-1 . For parameters ( α , β ) (\alpha,\beta) satisfying either α ≥ β ≥ 1 / 2 \alpha\geq\beta\geq 1/2 or α ≥ 1 / 2 \alpha\geq 1/2 , β = − 1 / 2 \beta=-1/2 , we prove the inequality P ⁡ ( 1 ) − P ⁡ ( x ) ≥ P ′ ​ ( x ) ​ ( 1 − x ) , x ∈ [ 0 , 1 ] , P(1)-P(x)\geq P^{\prime}(x)\,(1-x),\quad x\in[0,1], which may be viewed as a complement to Finite Increment Theorem for Jacobi polynomials. Keywords: Finite Increment Theorem, Chebyshev polynomials, Jacobi polynomials, nonnegative expansion. 2020 Mathematics Subject Classification: 41A17. 1. Introduction and statement of the result In [ 3 ] we investigated two trigonometric inequalities proposed by M. S. Robertson [ 5 ] and R. Askey and G. Gasper [ 2 ] , respectively (see (1.29) and (8.17) in [ 1 ] ). These inequalities were restated in [ 3 ] in terms of the Chebyshev polynomial of the first kind T_{n}(x)=\cos(n\theta),\quad x=\cos\theta\,, and the following refinement was proved: Theorem A. Let n\in\mathbb{N} , n\geq 4 . Then (1) T_{n}(x)+2-\frac{x+2}{n^{2}}\,T_{n}^{\prime}(x)-a(n)\frac{1-x}{n^{2}}\,\big(n^{2}-T_{n}^{\prime}(x)\big)\geq 0,\quad x\in[-1,1], where (2) a(n)=\begin{cases}\frac{1}{1+\cos\frac{\pi}{n}}\,,&\textit{ if }\ n\ \textit{ is even},\\ \frac{1}{1+\cos\frac{2\pi}{n}}\,,&\textit{ if }\ n\ \textit{ is odd.}\end{cases} The constant a(n) in ( 2 ) is the best possible in the sense that ( 1 ) fails for any larger constant. The equality in ( 1 ) is attained only at x=1 and x=-\cos\frac{\pi}{n} if n is even, and at x=\pm 1 and x=-\cos\frac{2\pi}{n} if n is odd. The inequalities of Robertson and of Askey and Gasper are obtained with choosing in ( 1 ) a(n)=0 and a(n)=1/2 , respectively. We note that Theorem A is true also for n=1,\,2,\,3 , in which cases the inequality in ( 1 ) becomes an identity. For x\in[0,1] the inequality in ( 1 ) holds true with a larger constant a(n) , namely with a=1 , yielding (cf. [ 3 , Theorem 2] ) Theorem B. For every n\in\mathbb{N} , n\geq 3 , the following inequality holds true: (3) T_{n}^{\prime}(1)-T_{n}^{\prime}(x)\geq(1-x)\,T_{n}^{\prime\prime}(x),\quad x\in[0,1]. The equality in ( 3 ) occurs only for x=1 and, if n\equiv 2\,(\!\!\!\mod 4) , for \,x=0 . In the cases n=0,\,1,\,2 , inequality ( 3 ) becomes an identity. The Chebyshev polynomials of the second kind \{U_{m}\} are related to the Chebyshev polynomial of the first kind by (4) U_{m}(x)=\frac{\sin(m+1)\theta}{\sin\theta}=\frac{T_{m+1}^{\prime}(x)}{m+1}\,,\quad x=\cos\theta. Therefore, inequality ( 3 ) can be reformulated as (5) U_{n}(1)-U_{n}(x)\geq(1-x)\,U_{n}^{\prime}(x),\quad n\in\mathbb{N}_{0},\ x\in[0,1]. According to the Finite Increment Theorem, \,U_{n}(1)-U_{n}(x)=(1-x)\,U_{n}^{\prime}(\xi) for some \xi\in(x,1) , hence inequality ( 3 ) may be viewed as a complement to this theorem applied to U_{n} . The Chebyshev polynomials of the first and the second kind are ultraspherical polynomials orthogonal in (-1,1) with respect to the weight functions 1/\sqrt{1-x^{2}} and \sqrt{1-x^{2}} , respectively. On the other hand, the ultraspherical (called also Gegenbauer) polynomials belong to the family of Jacobi polynomials \{P_{n}^{(\alpha,\beta)}(x)\}_{n\in\mathbb{N}_{0}} , which are orthogonal in [-1,1] with respect to the weight function w_{\alpha,\beta}(x)=(1-x)^{\alpha}(1+x)^{\beta} , \alpha,\,\beta>-1 . Inequality ( 5 ) says that for (\alpha,\beta)=(1/2,1/2) we have (6) P_{n}^{(\alpha,\beta)}(1)-P_{n}^{(\alpha,\beta)}(x)\geq(1-x)\,\frac{d}{dx}\big\{P_{n}^{(\alpha,\beta)}(x)\big\}\,,\quad x\in[0,1]\,. A rather natural question is: for which other parameters (\alpha,\beta) the inequality ( 6 ) holds true ? The aim of this note is to provide a partial answer of the above question. Our result is Theorem 1 . The inequality ( 6 ) holds true in the following cases: (i) \alpha\geq\frac{1}{2},\ \beta=-\frac{1}{2}\, ; (ii) \alpha\geq\beta\geq\frac{1}{2}\, . Unless \alpha=\beta=\frac{1}{2} , for n\geq 3 the equality in ( 6 ) is attained only at x=1 . The special case \alpha=\frac{1}{2} , \beta=-\frac{1}{2} of Theorem 1 (i) comes down to a trigonometric inequality, which to the best of our knowledge is new: Corollary 2 . For every n\in\mathbb{N},\;n\geq 2 , the following inequality holds true: 2n+1-(n+2)\,\frac{\sin n\theta}{\sin\theta}+(n-1)\,\frac{\sin(n+1)\theta}{\sin\theta}\geq 0\,,\quad\theta\in\big[0,\frac{\pi}{2}\big]\,. The equality is attained only for \theta=0 . The proofs are given in the next section. We provide a short proof of Theorem B for two reasons: firstly, for the readers convenience, and secondly, because Theorem B constitutes the main ingredient of our arguments. 2. Proofs 2.1. Proof of Theorem B Denote by \tau the largest zero of T_{n}^{\prime\prime} , then \cos\frac{2\pi}{n}<\tau<\cos\frac{\pi}{n}\,. T_{n}^{\prime\prime}(x) increases monotonically in [\tau,1] , therefore T_{n}^{\prime}(1)-T_{n}^{\prime}(x)=\int\limits_{x}^{1}T_{n}^{\prime\prime}(u)\,du\geq(1-x)T_{n}^{\prime\prime}(x),\quad x\in[\tau,1], with the equality only when x=1 . Now the assertion of Theorem B obviously follows when n=3 , and we assume in what follows n\geq 4 . Since the left-hand side of ( 3 ) is positive in [0,1) while \,T_{n}^{\prime\prime}(x)<0 for x\in\big[\cos\frac{2\pi}{n},\tau\big) , it follows that ( 3 ) holds true with ">" sign when x\in\big[\cos\frac{2\pi}{n},\tau\big) . Thus, we need to prove the inequality in ( 3 ) for x\in[0,\cos\frac{2\pi}{n}) and n\geq 5 . Let us set (7) \mathcal{L}(x)=y^{\prime}(1)-y^{\prime}(x)-(1-x)y^{\prime\prime}(x)\,,\qquad y=T_{n}(x)\,. By using the differential equation for y (see, e.g., [ 4 , Eqn. (1.92)] ) (8) (1-x^{2})y^{\prime\prime}(x)-x\,y^{\prime}(x)+n^{2}y(x)=0 and the representations T_{n}(x)=\cos n\theta,\ T_{n}^{\prime}(x)=n\,\frac{\sin n\theta}{\sin\theta},\quad x=\cos\theta, we find \mathcal{L}(0)=\begin{cases}2n^{2},&\text{ if }n\equiv 0\,(\!\!\!\!\mod 4),\\ n^{2}-n,&\text{ if }n\equiv 1\,(\!\!\!\!\mod 4),\\ 0,&\text{ if }n\equiv 2\,(\!\!\!\!\mod 4),\\ n^{2}+n,&\text{ if }n\equiv 3\,(\!\!\!\!\mod 4).\end{cases} Thus, \mathcal{L}(0)\geq 0 and the equality is attained only when n=4k+2,\ k\in\mathbb{N} . It remains to prove that \mathcal{L}(x)>0 for x\in(0,\cos\frac{2\pi}{n}) , and it suffices to verify this inequality at the points of local extremum of \mathcal{L}(x) therein. Since \mathcal{L}^{\prime}(x)=(1-x)y^{\prime\prime\prime}(x) , the latter points are the zeros of y^{\prime\prime\prime}(x) in \,(0,\cos\frac{2\pi}{n}) . Let t\in(0,\cos\frac{2\pi}{n})\, be a zero of y^{\prime\prime\prime}(x) . Note that \cos\frac{2\pi}{n}=1-2\sin^{2}\frac{\pi}{n}<1-2\Big(\frac{2}{\pi}\frac{\pi}{n}\Big)^{2}=1-\frac{8}{n^{2}}\,, therefore 0<t<1-\frac{8}{n^{2}}\,. Differentiating ( 8 ) at x=t and using y^{\prime\prime\prime}(t)=0 , we express y^{\prime\prime}(t) through y^{\prime}(t) as y^{\prime\prime}(t)=\frac{n^{2}-1}{3t}\,y^{\prime}(t). By plugging this expression in equation ( 8 ) with x=t , we obtain (9) y^{\prime}(t)=-\frac{3n^{2}t}{n^{2}-1-(n^{2}+2)t^{2}}\,y(t)\,, and consequently (10) y^{\prime\prime}(t)=-\frac{n^{2}(n^{2}-1)}{n^{2}-1-(n^{2}+2)t^{2}}\,y(t)\,. Now we replace y^{\prime}(1)=n^{2} and the expressions from ( 9 ) and ( 10 ) in \mathcal{L}(t) to obtain (11) \mathcal{L}(t)=n^{2}\Big[1+\frac{n^{2}-1-(n^{2}-4)t}{n^{2}-1-(n^{2}+2)t^{2}}\,y(t)\Big]\,. The quotient in front of y(t) in the last expression is positive. Indeed, this is obvious for the numerator, and for the denominator follows from t^{2}<t<1-\frac{8}{n^{2}}<1-\frac{3}{n^{2}+2}=\frac{n^{2}-1}{n^{2}+2}\,. Since y(t)\geq-1 , we conclude from ( 11 ) that \mathcal{L}(t)\geq n^{2}\Big[1-\frac{n^{2}-1-(n^{2}-4)t}{n^{2}-1-(n^{2}+2)t^{2}}\Big]=\frac{t\big[n^{2}-4-(n^{2}+2)t)\big]}{n^{2}-1-(n^{2}+2)t^{2}}\,. The numerator of the last quotient is positive, which is seen from t<1-\frac{8}{n^{2}}<1-\frac{6}{n^{2}+2}=\frac{n^{2}-4}{n^{2}+2}\,. Hence, \mathcal{L}(t)>0 and the proof of Theorem B is accomplished. 2.2. Proof of Theorem 1 We start with the special case \,(\alpha,\beta)=\big(\frac{1}{2},-\frac{1}{2}\big) of Theorem 1 (i). Apart from a constant multiplier, the Jacobi polynomial P_{n}^{(1/2,-1/2)}(x) coincides with the Chebyshev polynomial of the fourth kind (12) W_{n}(x)=\frac{\sin\big(n+\frac{1}{2}\big)\theta}{\sin\frac{\theta}{2}},\qquad x=\cos\theta\,. This representation and ( 4 ) imply (13) U_{n}(x)+U_{n-1}(x)=\frac{\sin(n+1)\theta}{\sin\theta}+\frac{\sin n\theta}{\sin\theta}=\frac{\sin\big(n+\frac{1}{2}\big)\theta}{\sin\frac{\theta}{2}}=W_{n}(x)\,. Now using ( 5 ), we obtain for x\in[0,1] \begin{split}W_{n}(1)-W_{n}(x)&=U_{n}(1)-U_{n}(x)+U_{n-1}(1)-U_{n-1}(x)\\ &\geq(1-x)U_{n}^{\prime}(x)+(1-x)U_{n-1}^{\prime}(x)\\ &=(1-x)W_{n}^{\prime}(x)\geq 0\,.\end{split} For n\geq 3 the equality occurs only for x=1 . Indeed, according to Theorem B, the equality can only occur at x=0 or x=1 , but it is not possible for both U_{n}(1)-U_{n}(x)-(1-x)U_{n}^{\prime}(x) and U_{n-1}(1)-U_{n-1}(x)-(1-x)U_{n-1}^{\prime}(x) to vanish at x=0 . We shall need the next lemma, which provides a positive expansion property of Jacobi

Read the complete text