三角不等式

Let $n$ be a natural number and let $0\lt x\lt{\pi}$. Then, here are my questions.

Question 1: Is the following true?
$$\sum_{k=1}^{n}\frac{\cos(kx)}{k}\gt -1$$

Question 2: Is the following true?
$$\sum_{k=1}^{n}\frac{\sin(kx)}{k}\gt0$$



This is a possible hint for solution; perhaps someone can finish it along these lines (it won‘t fit as a comment). We have
$$\sin x+\dfrac{\sin 2x}{2}+\dfrac{\sin 3x}{3}+\ldots+ \dfrac{\sin nx}{n}=\sum_{k=1}^n\int_0^x\cos kt\,dt,$$
$$2\sum_{k=1}^n\cos kt=\sin((n+1/2)t)/\sin(t/2)-1$$ (by taking the real part of $\sum_{k=1}^n e^{ikt}$)
so we want to show
$$\int_0^x(\frac{\sin(n+1/2)t}{\sin(t/2)}-1)dt>0.$$
It‘s easy without that $-1$ (as $1/\sin (t/2)$ decreases); to do it really (with $-1$) a better estimate is needed.



TEOREMA (L.Fejer-1910, D. Jackson-1912, T.J.Gronwall- 1912). Pentru $x\in (0,\pi)$ au loc inegalitatile
\[(1)\; \; \; \; \; \; \begin{array}{|c|}\hline \\ \sum_{k=1}^{n}\frac{\sin{kx}}{k}> 0 \\ \\ \hline \end{array}\; \; ,\; \; \forall n \in{\mathbb N}. \]

DEMONSTRATIA I .
Prin efectuarea unor calcule elementare se constata
\[\frac{d}{d\varphi}\left\{ \frac{\sin{2k\phi}}{(\sin{\phi})^{2k}}\right\}=-2k\frac{\sin{(2k-1)\phi}}{(\sin{\phi})^{2k+1}}\; . \]
Efectuand calculele precum si substitutia $\phi \leadsto \frac{x}{2}\; ,\;$ prin integrare avem
\[\frac{\sin{kx}}{k}=2\left(\sin{\frac{x}{2}}\right)^{2k}\int_{\frac{x}{2}}^{\frac{\pi}{2}}\frac{\sin{(2k-1)\phi}}{(\sin \phi)^{2k+1}}\; d \phi \; \; ,\; k\in{\mathbb N}. \]
Insumand pentru $k\in \{1,2,...,n\}$ se obtine
\[\sum_{k=1}^{n}\frac{\sin{kx}}{k}=2\int_{\frac{x}{2}}^{\frac{\pi}{2}}\sum_{k=1}^{n}\left[r(x,\phi)\right]^{k}\frac{\sin{(2k-1)\phi }}{\sin{\phi}}\; d\phi \]
unde \[r(x,\phi): =\left(\frac{ \sin\frac{x}{2}}{\sin{\phi}}\right)^{2}\in [0,1] \]
pentru $0\le \frac{x}{2}\le \phi<\frac{\pi}{2}\; .$ Egalitatea de mai sus implica
\[\sum_{k=1}^{n}\frac{\sin{kx}}{k}=\int_{x}^{\pi}\sum_{k=1}^{n}\left[r(x,\frac{\psi}{2})\right]^{k}\frac{\sin{\left(k-\frac{1}{2}\right)\psi }}{\sin{\frac{\psi}{2}}}\; d\psi\; . \]
Aplicand identitatea lui Abel ("‘insumarea prin parti"), adica
\[\sum_{k=1}^{n}A_{k}B_{k}=\sum_{k=1}^{n-1}\left(A_{k}-A_{k+1}\right)\sum_{j=1}^{k}B_{j}+A_{n}\sum_{j=1}^{n}B_{j}\; , \]
si notand
\[\left\{\begin{array}{rcl}{\mathbf r}&: =&r(x,\frac{\psi}{2})\; \; , \; \;{\mathbf r}\in [0,1]\; \; \; \mbox{vezi}\; \; ()\\{\mathcal F}_{n}(\psi)&: =&\sum_{k=1}^{n}\sin{\left(k-\frac{1}{2}\right)\psi}\; \; \; , \; \;{\mathcal F}_{n}(\psi)\ge 0 \; \; \mbox{dac‘a}\; x\in [0,\pi]\; \;-\mbox{vezi L. Fej\‘er}\end{array}\right.\; , \] din identitatea lui Abel g‘asim
\[\sum_{k=1}^{n}\frac{\sin{kx}}{k}=\int_{x}^{\pi}\left\{(1-{\mathbf r})\sum_{k=1}^{n-1}{\mathbf r}^{k-1}{\mathcal F}_{k}(\psi)+{\mathbf r}^{n}{\mathcal F}_{n}(\psi)\right\}\frac{d\psi}{\sin{\frac{\psi}{2}}}\; . \]
Deoarece ${\mathcal F}_{n}(\psi)=\frac{1-\cos(n\psi)}{2\sin{\frac{\psi}{2}}}\; ,$ concludem cu
\[\begin{array}{|c|}\hline \\ \sum_{k=1}^{n}\frac{\sin{kx}}{k}=\frac{1}{2}\int_{x}^{\pi}\left\{(1-{\mathbf r})\sum_{k=1}^{n-1}{\mathbf r}^{k-1}\left(1-\cos(k\psi)\right)+{\mathbf r}^{n}(1-\cos(n\psi))\right\}\frac{d\psi}{\left(\sin{\frac{\psi}{2}}\right)^{2}}\\ \\ \hline \end{array}\; . \]
Aceasta reprezentare completeaza demonstra‘tia I.

DEMONSTRATIA II.
Fie $P_{n}(x)$ polinomul lui Legendre de gradul $n$, adic‘a
\[\begin{array}{c}P_{n}(x)=\frac{1}{n! 2^{n}}\left[(x^{2}-1)^{n}\right)^{(n)}={}_{2}F_{1}\left(-n,n+1;1;\frac{1-x}{2}\right)=\\ \\ =\frac{1}{2^{n}}\sum_{k=0}^{\lfloor \frac{n}{2}\rfloor}{n\choose 2k}(-1)^{k}\frac{(n+1)_{n-2k}(2k)!}{k!(n-k)!}x^{n-2k}\end{array}\]
unde
\[\begin{array}{c}{}_{2}F_{1}(-n;b;c;z): =\sum_{k=0}^{n}(-1)^{k}{n\choose k}\frac{(b)_{k}}{(c)_{k}}z^{k}\\ (d)_{k}: =d(d+1)\cdots (d+k-1)\; ,\; \; k\in{\mathbb N}\; ,\; (d)_{0}: =1 . \end{array}\]
Se cunosc urmatoarele:
-- radacinile lui $P_{n}(x)$ sunt reale,distincte, situate in $(-1,1)$;
-- $|P_{n}(t)|\le 1 \; ,\; \; \forall t\in [-1,1] .$
Demonstratia a II-a (vezi [16] precum si comentariile lui R.Askey )
se bazeaza pe identitatea:
\[\begin{array}{|c|}\hline \\ \sum_{k=1}^{n}\frac{\sin(k\cdot\arccos{x})}{k}= \frac{\sqrt{1-x}}{2}\int_{-1}^{x}\frac{1-P_{n}(y)}{1-y}\frac{dy}{\sqrt{x-y}}\\ \\ \hline \end{array}\; ,\; x\in (-1,1). \; .\]

Observatii.

1) Inegalitatea (1) a fost conjecturata de catre Leopold (Lipot) Fejer. Ulterior a fost demonstrata de catre D.Jackson-[14] si T.H.Gronwall-[12].
In prezent se cunosc peste 100 de demonstratii. Am ales pentru a Va prezenta pe cele pe care subsemnatul le considera "mai simple".

Se spune ca pana la moasrte L.Fejer a cautat sa gaseasca noi demonstratii a lui (1). L.Fejer a predat si la Universitatea din Cluj (sub numele de L.Weiss).
2)Desi simpla la prima vedere, inegalitatea (1) a dat de furca matematicienilor.
3) Inegalitatea (1) intervine in urmatoarele domenii: Serii Fourier (fenomenul lui Gibbs - vezi [13]), Polinoame ortogonale ([1]-[5],[18] ), Functii Complexe (Demonstratia conjecturii lui Bieberbach,functii univalente, [4],[6]), Teoria Aproximarii ([5]).

BIBLIOGRAFIE.
[1]R. Askey , Orthogonal Polynomials and Special Functions,
Regional Conf.Lect.Appl.Math., vol.21,SIAM,Philadelphia,Pa., 1975.
[2]R.Askey Positive quadrature methods and positive polynomial sums, ‘in Approximation Theory V, Academic Press, 1986.
[3] R. Askey and J. Fitch , Integral reprezentations for Jacobi
polynomial amd some applications ,J.Math.Anal.Appl., 26 (1969)
411-437.
[4]R. Askey and G. Gasper , Positive Jacobi polynomial sums,(II),
Amer.J.Math., 98 (1976) 709-737.
[4]R. Askey and G. Gasper , Inequalities for polynomials}, ‘in "The Bieberbach Conjecture", Proc.of the Symposium on the Occasion
of the Proof, Mathematical Surveys and Monographs, 21, Amer.Mathematical Society, 1986, 7-32.
[5]H. Bavinck , Jacobi Series and Approximation,
Mathematical Centre Tracts 39, Mathematisch Centrum Amsterdam 1972.
[6] L. de Branges , The Story of the Verification of the Bieberbach Conjecture, ‘in " The Bieberbach Conjecture", Proc.of the Symposium on the Occasion of the Proof, Mathematical Surveys and Monographs, 21, Amer.Mathematical Society, 1986, 199-203.
[7] L. Fejer , Sur les fonctions bornees et integrables,
C.R.Acad.Sci.Paris 131 (1900) 984-987.
[8] L. Fejer , Sur le develpopment d‘une fonction arbitraire suivant les fonctions de Laplace, C.R.Acad.Sci.Paris , 146 (1908) 224-227.
[9]L. Fejer , Ueber die Laplacesche Reihe, Math. Ann. (1909)76-109.
[10] L. Fejer , Einige Saetze , die sich auf das Vorzeichen einer ganzen rationalen Funktion beziehen u.s.w., Monta.fuer Math. und Phys., 35 (1928) 305-344.
[11]L. Fejer , Gesammelte Arbeiten (I)-(II), Birkhauser Verlag, Basel, 1970 .
[12]T.H. Gronwall , Ueber die Gibbsche Erscheinung und die trigonometrischen Summen $\sin{x}+(1/2)\sin{2x}+...+(1/n)\sin{nx}$ , Math.Ann., 72 (1912) 228-243.
[13] E. Hewitt and R.E. Hewitt , The Gibbs-Wilbraham phenomenon: an epsiode in Fourier analysis, Arch.Hist.Exact Sci., 21 (1979) 129-160.
[14]D. Jackson , Ueber eine trigonometrische Summe, Rend.Circ.Mat.Palermo 32(1911) 257-262.
[15]A. Lupas , Advanced Problem 6517, Amer.Math.Monthly (1986) p. 305 ; (1988) p.264.
[16]A. Lupas , Advanced Problem 6585, Amer.Math.Monthly
(1988) p. 880 ; (1990) p.859-860.
[17]L. Lupas , An identity for ultraspherical polynomials, Revue d${}^{,}$Analyse numerique et de Theorie de l‘approximation, tome 24 , no.1-2 (1995) 181-185.
[18]G. Szego , Orthogonal Polynomials , Amer.Math.Soc.Colloq. Publications vol.23, fourth ed.,Amer.Math.Soc., Providence, R.I., 1975.



C?lcat-up inegalitate $ \sum_{k=1}^{n}\frac{\sin{kx}}{k}>x\left(1-\frac{x}{\pi}\right)^3$ de?ine pentru $ x \in (0,\pi).$

参考这里Sawtooth wave



**In short:** let $f_n(x)$ denote the function on the lhs of the inequality. Of course, $f_1(x)=\sin x\geq 0$ on $[0,\pi]$. We will prove that $f_n(x)\geq 0$ on $[0,\pi]$ by induction on $n$. It is not too hard to determine the local minima of $f_n$ on $[0,\pi]$ by investigating its derivative. Then Ma Ming observed that $f_n$ coincides with $f_{n-1}$ on these local minima. And the induction step follows easily. Of course, $f_n(0)=f_n(\pi)=0$. We will actually prove that
>$$
f_n(x)=\sum_{k=1}^n\frac{\sin kx}{k}>0\qquad\forall x\in(0,\pi).
$$

**Remark:** it is worth noting that the $f_n$‘s are the partial sums of the Fourier series of the same sawtooth function. Just [look at the case $n=6$][1], for instance, to see how they tend to approximate it nicely. [See here][2] to get an idea how to estimate the error in such approximations. As pointed out by math110, there are many proofs of this so-called Fejer-Jackson inequality. It can even be shown that the [$f_n$‘s are bounded below][3] by a certain nonnegative polynomial on $[0,\pi]$. The proof below is at the calculus I level. I‘m not sure it can be made more elementary.

**Proof:** first, $f_1(x)=\sin x$ is positive on $(0,\pi)$. Assume this holds for $f_{n-1}$ for some $n\geq 2$. Then observe that $f_n$ is differenbtiable on $\mathbb{R}$ with
$$
f_n‘(x)=\sum_{k=1}^n\cos kx=\mbox{Re} \sum_{k=1}^n (e^{ix})^k.
$$
For $x\in 2\pi \mathbb{Z}$, we have $f_n‘(x)=n$. So the zeros of $f_n‘$ are the zeros of
$$
\mbox{Re}\;e^{ix}\frac{e^{inx}-1}{e^{ix}-1}=\mbox{Re}\;e^{i(n+1)x/2}\frac{\sin (nx/2)}{\sin(x/2)}=\frac{\cos((n+1)x/2)\sin (nx/2)}{\sin(x/2)}.
$$
This yields
$$
\frac{nx}{2}\in \pi\mathbb{Z}\quad\mbox{or}\quad \frac{(n+1)x}{2}\in \frac{\pi}{2}+\pi\mathbb{Z}
$$
i.e.
$$
x\in \frac{2\pi}{n}\mathbb{Z}\quad\mbox{or}\quad x\in \frac{\pi}{n+1}+\frac{2\pi}{n+1}\mathbb{Z}.
$$
Between $0$ and $\pi$, these are ordered as follows:
$$
0<\frac{\pi}{n+1}<\frac{2\pi}{n}<\frac{3\pi}{n+1}<\frac{4\pi}{n}<\ldots < \frac{2\lfloor n/2\rfloor \pi}{n}\leq \pi.
$$
The sign of $f_n‘$ changes at each of these zeros, starting from a positive sign on $(0,\pi/(n+1))$. It follows that $f_n$ is positive on the latter, positive on the last interval (if nontrivial, i.e. in the odd case), with local minima at
$$\frac{2j\pi}{n}\qquad\mbox{for}\qquad j=1,\ldots,\lfloor n/2\rfloor.$$

But now here is Ma Ming‘s key observation: for these values, we have
$$
f_n\left(\frac{2j\pi}{n}\right)=f_{n-1}\left(\frac{2j\pi}{n}\right)+\sin\left(n\cdot\frac{2j\pi}{n}\right)=f_{n-1}\left(\frac{2j\pi}{n}\right)>0
$$
by induction step. It follows that $f_n(x)>0$ on $(0,\pi)$. QED.

[1]: http://www.wolframalpha.com/input/?i=sin%20%28x%29%2bsin%20%282x%29/2%2bsin%20%283x%29/3%2bsin%20%284x%29/4%2bsin%285x%29/5%2bsin%286x%29/6
[2]: https://math.stackexchange.com/questions/57054/asymptotic-error-of-fourier-series-partial-sum-of-sawtooth-function
[3]: https://math.stackexchange.com/questions/177995/a-pseudo-fejer-jackson-inequality-problem

参考:论坛1论坛1.1这里

论坛2

论坛3



How to prove that $\forall x \in \mathbb{R}$, $n \in \mathbb{N}$, we have

\begin{align}
\sum_{k=1}^{n}\frac{|\sin{kx}|}{k}\ge |\sin{nx}| ?
\end{align}

Given $n\ge 1$ and $x\in \Bbb R$, denote
$$f_n(x)=\sum_{k=1}^n\frac{|\sin kx|}{k} \quad\text{and}\quad g_n(x)=f_n(x)-|\sin nx|.$$

----------

> **Lemma:** For every $n\ge 1$,
>
> (i) $f_n$ is increasing on $[0,\frac{\pi}{n+1}]$;
> (ii) $g_n$ is increasing on $[0,\frac{\pi}n]$;
> (iii) $f_n\ge 1$ on $[\frac{\pi}{2n},\frac{\pi}{2}]$.

**Proof of Lemma:** Note that when $x\in [0,\frac{\pi}n]$,
$$f_n(x)=\sum_{k=1}^n\frac{\sin kx}{k},\quad\text{and}\quad g_n(x)=f_n(x)-\sin nx,$$
so
$$f_n‘(x)=\sum_{k=1}^n \cos kx \quad\text{and}\quad g_n‘(x)=f_n‘(x)-n\cos nx.$$
(i) Given $x\in[0,\frac{\pi}{n+1}]$, noting that $\cos\frac{kx}{2}\ge 0$ for $k=0,\pm1,\dots, \pm (n+1)$, we have
$$f_n‘(x)=\sum_{k=1}^n \frac{\cos kx +\cos(n+1-k)x}{2}=\cos \frac{(n+1)x}{2} \cdot\sum_{k=1}^n \cos\frac{(n+1-2k)x}{2}\ge 0.$$

(ii) Since the cosine function is decreasing on $[0,\pi]$, when $x\in [0,\frac{\pi}{n}]$, $\cos k x\ge \cos nx$ for $k=1,\dots,n$, so $g‘(x)\ge 0$.

(iii) When $n=1$, the statement is clearly true; when $n=2$, since $f_2$ is concave on $[\frac{\pi}{4},\frac{\pi}{2}]$, $f_2(\frac{\pi}{4})>1$ and $f_2(\frac{\pi}{2})=1$, the statement is also true. By induction, we may assume that $f_{n-1}\ge 1$ on $[\frac{\pi}{2(n-1)},\frac{\pi}{2}]$ for some $n \ge 3$, and the conclusion $f_n\ge 1$ on $[\frac{\pi}{2n},\frac{\pi}{2}]$ follows from the facts below. Firstly, $f_n\ge f_{n-1}$; secondly, $f_n$ is increasing on $[0,\frac{\pi}{n+1}]\supset [\frac{\pi}{2n},\frac{\pi}{2(n-1)}]$; thirdly,
$$\sin \frac{\pi t}{2}\ge t,\ \forall t\in[0,1]\Longrightarrow f_n(\frac{\pi}{2n})=\sum_{k=1}^n\frac{\sin \frac{k\pi}{2n}}{k}\ge 1.\qquad \square$$

----------
Now we can prove that $g_n\ge 0$ by using the lemma. Since $g_n(\pi\pm x)=g_n(x)$, we may focus on $x\in[0,\frac{\pi}{2}]$. Since $g_n(0)=0$, by (ii), we know that $g_n(x)\ge 0$ on $[0,\frac{\pi}{n}]$. Since $g_n\ge f_n -1$, by (iii) we know that $g_n\ge 0$ on $[\frac{\pi}{2n},\frac{\pi}{2}]$.

参考:这里

原文地址:https://www.cnblogs.com/Eufisky/p/8995608.html

时间: 2024-10-29 18:00:30

三角不等式的相关文章

关于2-范数三角不等式的证明

数值计算课上的作业,回去想了一些,偶然看到豆瓣上11年有同学也问了,看了评论有了思路,可以用柯西不等式.sqrt((x1+y1)^2 + ...+(xn+yn)^2)=sqrt(x1^2 +...+xn^2+y1^2+...+yn^2+2*x1*y1+...+2*xn*yn)<=sqrt(x1^2+...+xn^2+y1^2+...+yn^2+?2*sqrt(x1^2+...+xn^2)*sqrt(y1^2+...+yn^2))=sqrt(x1^2+...xn^2)+sqrt(y1^2+...+

各种不等式的解法收集【初级辅导和中级辅导】

高三数学学习的第一课----不等式的解法 1.一元二次不等式 1.1 数字系数的一元二次不等式,如\(x^2<3\),\(x^2+2x<0\),\(x^2-3x+2<0\) 1.2 字母系数的一元二次不等式,如\(x^2-(a+a^2)x+a^3<0.(a\neq0)\) 1.3 能转化为一元二次不等式,如\((x^2-3x+2)\cdot(x+1)<0,2^{x^2-x}<4\), 如果能理解不等式中的\(x\)的内涵,\(x\Rightarrow 代数式\),则可以

代数不等式与超越不等式

前言 超越不等式 如果不等式的两边至少有一个是超越函数,则称这个不等式为超越不等式.如\(2^x>x-1\),包括指数不等式.对数不等式.三角不等式和反三角不等式等. 备注:代数函数1:超越函数2:代数不等式3: 超越不等式求解思路 例1求解关于\(x\)的不等式\((2^x)-3\cdot 2^x+2<0\): 分析:换元法,令\(2^x=t>0\),则原超越不等式可以等价转化为代数不等式,不过是带有条件的,比如\(t>0\); 转化为\(t^2-3t+2<0(t>0

浅谈Cauchy不等式

形式 \[ \sum_{i=1}^{n}a_i^2 \sum_{i=1}^{n}b_i^2 \geq \sum_{i=1}^{n}a_i^{2}b_i^2 \] 等号成立的条件: \[ iff:b_i=0 || \exists k \in \mathbb {R},a_i=k \cdot b_i(i \in \mathbb{N^+}) \] 证明 法一:参数配方 思路:巧妙的把常数与方程结合起来,利用性质即可. 证明: 构造函数: \[ f(t)=\sum_{i=1}^{n}b_i^2\cdot

Clustering by Passing Messages Between Data Points(Brendan J.Frey* and Delbert Dueck)例子

例1,数据点聚类:AP应用到25个二维数据中,使用负平法误差作为相似度  聚类数目不用预先指定 AP的一个优点是聚类数目不用预先指定,而是在消息传递方法中慢慢浮现,取决于输入参考度(preference),这种自动模型的选择,基于先验指定每一个数据点有多合适作为exemplar. 下图表示输入常量参考度对聚类数目的影响,这种关系近乎同样于在精确最小平方误差中的关系. 例2,人脸识别:使用优化标准为均方差,识别900张灰度图像   AP一致的能够实现更低的误差,在时间上花费要少于两个数量级. AP

算法导论——lec 13 贪心算法与图上算法

之前我们介绍了用动态规划的方法来解决一些最优化的问题.但对于有些最优化问题来说,用动态规划就是"高射炮打蚊子",采用一些更加简单有效的方法就可以解决.贪心算法就是其中之一.贪心算法是使所做的选择看起来是当前最佳的,期望通过所做的局部最优选择来产生一个全局最优解. 一. 活动选择问题 [问题]对几个互相竞争的活动进行调度:活动集合S = {a1, a2, ..., an},它们都要求以独占的方式使用某一公共资源(如教室),每个活动ai有一个开始时间si和结束时间fi ,且0 ≤ si &

Clustering by fast search and find of desity peaks(基于快速搜索与寻找密度峰值的聚类)

基于快速搜索与寻找密度峰值的聚类(Alex Rodriguez and Alessandro Laio) 摘要:聚类分析目的是基于元素之间的相似度对其进行分类,应用范围从天文学到生物信息学.文献计量学到模式识别.我们提出一种方法,思想基于簇中心具有比其邻居更大密度的特点以及与更大密度点之间有一个相对较大的距离(1.簇中心点有相对高的密度 2.簇中心点之间距离一般较大,即不同类别之间一般距离较远),这种思想形成了簇数目直观出现的聚类机制的基础,自动发现和排除异常点,同时在识别簇时,不用关心其形状和

bzoj1820

想出如何记录状态后就不难了,转移很好想,但也许略复杂... dp(i,j,k)表示在处理第i个业务,另外2个在j,k处,转移的话就是可以移动a[i-1].j或k,第一维可以滚动 1 #include<cstdio> 2 #include<cstdlib> 3 #include<cstring> 4 #include<ctime> 5 #include<cmath> 6 #include<iostream> 7 #include<

poj3259(spfa)

自己的第一道spfa,纪念一下,顺便转载一下spfa的原理.先po代码: #include <iostream> #include <queue> using namespace std; const int MAX = 999999; const int MAXN = 501; int minimum(int a, int b){ return a > b ? b : a; } int main() { int t; cin >> t; while (t--){