Further results on \(L^{1}\)-convergence of some modified complex trigonometric sums

Xhevat Z. Krasniqi\(^\ast \)

December 20, 2015.

\(^\ast \)Department of Mathematics and Informatics, Faculty of Education, University of Prishtina "Hasan Prishtina", Avenue "Mother Theresa" 5, 10000 Prishtina, Kosovo, e-mail: xhevat.krasniqi@uni-pr.edu.

In this paper we have defined a new class of numerical sequences, which tend to zero, briefly denoted by \(\mathbb {K}^{2}\). Moreover, employing such class of numerical sequences we have studied \(L^{1}\)-convergence of some modified complex trigonometric sums introduced previously by others.

MSC. 42A20, 42A32.

Keywords. \(L^{1}\)-convergence, null sequence, trigonometric series, modified sums.

1 Introduction

Let \((c_{k})\), \(k\in \{ 0,\pm 1,\pm 2,\dots \} \), be a sequence of complex numbers and let

\begin{equation} \label{eq11} \sum _{k=-\infty }^{\infty }c_{k}e^{ikx} \end{equation}
1

be a formal complex trigonometric series with its partial sums

\begin{equation} \label{eq12} S_{n}(x)=\sum _{k=-n}^{n}c_{k}e^{ikx}, \quad n\in \{ 0,1,2,\dots \} . \end{equation}
2

Let

\[ \| f\| _{L^{1}}=\frac{1}{2\pi }\int _{-\pi }^{\pi }|f(x)|dx \]

be the \(L^{1}\)-norm of a function \(f\in L^1\).

The following interesting statement is a well-known one: If a trigonometric series converges in \(L^1\)-norm to a function \(f\in L^1\), then it is the Fourier series of the function \(f\). Riesz (see [ 5 ] , Vol. II, Chap. VIII, \(\S \) 22) gave a counter example showing that in the metric \(L^1\) we can not expect the converse of above mention statement to be true. This fact motivated the various authors to study the \(L^{1}\)-convergence of trigonometric series, introducing the so-called modified cosine and sine sums, since these modified sums approximate their limits better than the classical trigonometric series in the sense that they converge in \(L^1\)-norm to the sum of the trigonometric series whereas the classical series itself may not. C. S. Rees and C. V. Stanojevic [ 3 ] for the first time introduced the following type of modified cosine sums

\[ f_{n}(x)=\tfrac {1}{2}\sum _{k=0}^{n}\triangle a_{k}+\sum _{k=1}^{n}\sum _{j=k}^{n}\triangle a_{j}\cos kx , \]

and obtained a necessary and sufficient condition for the integrability of the limit of these sums, where \(\triangle a_{j}=a_{j}-a_{j+1}\).

Then several interesting properties (their integrability [ 2 ] or \(L^{1}\)-convergence [ 15 ] ) of these sums were investigated imposing several conditions on the coefficients \(a_{k}\) in "old papers" [ 1 ] , [ 2 ] , [ 6 ] or in some "new papers" [ 4 ] , [ 9 ] , [ 13 ] , [ 14 ] , [ 16 ] .

After introducing the sums \(f_{n}(x)\), B. Ram and S. Kumari [ 8 ] seems to be motivated to introduce the set of the sums

\[ h_{n}(x)=\frac{a_0}{2}+\sum _{k=1}^{n}\sum _{j=k}^{n}\triangle \left(\frac{a_{j}}{j}\right)k\cos kx \]
\[ g_{n}(x)=\sum _{k=1}^{n}\sum _{j=k}^{n}\triangle \left(\frac{a_{j}}{j}\right)k\sin kx \]

and studied their \(L^{1}\)-convergence under condition that the coefficients \(a_n\) belong to the class \(\mathbb {R}\).

Definition 1 [ 12 ]

If \(a_{k}\to 0\) as \(k\to \infty \) and

\[ \sum _{k=1}^{\infty }k^{2}\left|\triangle ^{2}\left(\frac{a_{k}}{k}\right)\right|{\lt}\infty \]

then it is said that \((a_{k})\) belongs to the class \(\mathbb {R}\), where \(\triangle ^{2}a_{j}=a_{j}-2a_{j+1}+a_{j+2}\).

The above definition were introduced by T. Kano who verified a result which we will formulate it as follows.
Theorem 2 [ 12 ]

If \((a_{k})\in \mathbb {R}\) then the series (1) and (2) are Fourier series or equivalently they represent integrable functions.

Among others, using this result B. Ram and S. Kumari proved the following result.
Theorem 3 [ 8 ]

Let \((a_{k})\in \mathbb {R}\). Then for \(x\in (0,\pi ]\)

\begin{equation} \label{eq13} \lim _{n\to \infty }t_{n}(x)=t(x),\quad t\in L(0,\pi ] , \end{equation}
3

and

\begin{equation} \label{eq14} \| t_{n}-t\| _{L^1}\to 0,\, \, \text{as}\, \, n\to \infty , \end{equation}
4

where \(t_{n}(x)\) represents either \(h_{n}(x)\) or \(g_{n}(x)\).

The authors of [ 9 ] have introduced the following modified complex trigonometric sums

\begin{align*} g_{n}^{c}(x)& =S_{n}(x)+\frac{i}{2\sin x}[ c_{n}e^{i(n+1)x}- c_{-n}e^{-i(n+1)x}+c_{n+1}e^{inx}- c_{-(n+1)}e^{-inx}\\ & \quad +(c_n-c_{n+2})E_{n}(x)+(c_{-(n+2)}-c_{-n})E_{-n}(x)], \end{align*}

where \(E_{n}(x)=\sum _{j=0}^{n}e^{ijx}\).

Note that the above sums are indeed the complex form of the modified sine and cosine sums introduced in [ 10 ] and [ 11 ] respectively.

Also they have introduced the following definition.

Definition 4

A sequence \((c_{k})\) of complex numbers belongs to class \(J^{*}\) if \(\lim _{k\to \infty }c_k=0\), and there exists a sequence \((A_{k})\) such that

\begin{equation} \label{eq15} A_{k}\downarrow 0, \quad \text{as}\quad k\to \infty , \end{equation}
5

\begin{equation} \label{eq16} \sum _{k=1}^{\infty }kA_{k}< \infty , \end{equation}
6

and

\begin{equation} \label{eq17}\left|\triangle \left(\frac{c_{k}-c_{-k}}{k}\right)\right|\leq \frac{A_{k}}{k}, \quad \forall k. \end{equation}
7

Throughout this paper we will denote by \(S_{n}(x)\) the partial sums of the series (1) and \(\lim _{n\to \infty }S_{n}(x)=f(x)\).
Theorem 5 [ 8 ]

Let \((c_{k})\) belongs to the class \(J^{*}\) then

\begin{equation} \label{eq18} \lim _{n\to \infty }g_{n}^{c}(x)=f(x),\quad \text{exists for}\quad |x|\in (0,\pi ], \end{equation}
8

\begin{equation} \label{eq19} f\in L^{1}(0,\pi ]\quad \text{and}\quad \| g_{n}^{c}-f\| _{L^{1}}\to 0,\, \, \text{as}\, \, n\to \infty , \end{equation}
9

and

\begin{equation} \label{eq20} \| S_{n}-f\| _{L^1}\to 0,\, \, \text{as}\, \, n\to \infty . \end{equation}
10

Here and in sequel we will use the notation \(\bigtriangleup ^{2}b_{k}=\bigtriangleup b_{k}-\bigtriangleup b_{k+1}, k\in \{ 1,2,\dots \} \). Now we are going to introduce the following class of sequences of complex numbers.

Definition 6

A sequence \((c_{k})\) of complex numbers belongs to class \(\mathbb {K}^{2}\) if \(\lim _{k\to \infty }c_k=0\), and there exists a sequence \((A_{k})\) such that

\begin{equation} \label{eq15} A_{k}\downarrow 0, \quad \text{as}\quad k\to \infty , \end{equation}
11

\begin{equation} \label{eq16} \sum _{k=1}^{\infty }k^{2}A_{k}< \infty , \end{equation}
12

and

\begin{equation} \label{eq17}\max \left\{ \left|\triangle ^2\left(\frac{c_{k}}{k}\right)\right| , \left|\triangle ^2\left(\frac{c_{-k}}{k}\right)\right|\right\} \leq \frac{A_{k}}{k^2},\quad k\in \{ 1,2,\dots \} . \end{equation}
13

Next example shows that the there exist sequences that belong or not belong to the class \(\mathbb {K}^{2}\).

Let \(\left(c_k\right)\) be a sequence defined by its general term \(c_k:=\frac{1}{n^2}\) , \(n\in \{ 1,2,\dots \} \). Then, \(\left|\bigtriangleup ^2\left(\frac{c_{\pm k}}{k}\right)\right|\leq \frac{4}{k^3}=\frac{A_{k}}{k^2}\), \(A_{k}=\frac{4}{k}\downarrow 0\), and \(\sum _{k=1}^{\infty }k^2A_{k}=+\infty \), which means that \(\left(c_k\right)\not\in \mathbb {K}^2\).

On the other hand, let \(\left(\overline{c}_k\right)\) be a sequence defined by its general term \(\overline{c}_k=\frac{1}{n^5}\) , \(n\in \{ 1,2,\dots \} \). Then, \(\left|\bigtriangleup ^2\left(\frac{c_{\pm k}}{k}\right)\right|\leq \frac{4}{k^6}=\frac{A_{k}}{k^2}\), \(A_{k}=\frac{4}{k^4}\downarrow 0\), and \(\sum _{k=1}^{\infty }k^2A_{k}{\lt}+\infty \), which means that \(\left(\overline{c}_k\right)\in \mathbb {K}^2\).

The aim of this paper is to study \(L^{1}\)-convergence of sums \(g^{c}_{n}(x)\) under condition that the coefficients \(c_k\) belong to the class \(\mathbb {K}^{2}\).

Closing this section, we recall the well-known equality named as Abel’s transformation: Let \(n\) be a positive integer, and \(a_{1}, a_2,\dots , a_{n}\) and \(b_{1}, b_2,\dots , b_{n}\) be two sequences. If \(S_k=a_1+a_2+\cdots +a_n\), then

\[ \sum _{k=1}^{n}a_kb_k=\sum _{k=1}^{n-1}S_k(b_k-b_{k+1})+S_nb_n. \]

2 Helpful Lemmas

Let

\[ \widetilde{D}_{n}(x)=\sum _{j=1}^{n}\cos (jx)=\frac{\sin \left(n+\frac{1}{2}\right)x}{2\sin \frac{x}{2}} \]

and

\[ \widetilde{D}_{n}(x)=\sum _{j=1}^{n}\sin (jx)=\frac{\cos \frac{x}{2}-\cos \left(n+\frac{1}{2}\right)x}{2\sin \frac{x}{2}} \]

be the Dirichlet’s and conjugate Dirichlet’s kernels respectively.

The following statements are needed for the proof of the main result.

Lemma 7 [ 7 ]

Let \(r\) be a non-negative integer and \(0{\lt}\varepsilon {\lt}\pi \). Then there exists \(M_{r\varepsilon }{\gt}0\) such that for all \(\varepsilon \leq |x|\leq \pi \) and all \(n\geq 1\),

  1. \(|E_{n}^{(r)}(x)|\leq \frac{M_{r\varepsilon }n^r}{|x|}\),

  2. \(|E_{-n}^{(r)}(x)|\leq \frac{M_{r\varepsilon }n^r}{|x|}\),

  3. \(|D_{n}^{(r)}(x)|\leq \frac{2M_{r\varepsilon }n^r}{|x|}\),

  4. \(|\widetilde{D}_{n}^{(r)}(x)|\leq \frac{M_{r\varepsilon }n^r}{|x|}\),

Lemma 8 [ 9 ]

For \(n\geq 1\), we have

  1. \(\left\| \frac{E_{n}(x)}{2\sin x}\right\| _{L^1}=o(n)\) as \(n\to \infty \),

  2. \(\left\| \frac{E_{-n}(x)}{2\sin x}\right\| _{L^1}=o(n)\) as \(n\to \infty \),

  3. \(\left\| \frac{e^{inx}}{2\sin x}\right\| _{L^1}=o(\log n)\) as \(n\to \infty \).

Lemma 9

Let \(r\) be a non-negative integer and \(0{\lt}\varepsilon {\lt}\pi \). Then there exists \(M_{r\varepsilon }{\gt}0\) such that for all \(\varepsilon \leq |x|\leq \pi \) and all \(n\geq 1\),

  1. \(|\overline{E}_{n}'(x)|\leq \frac{M_{r\varepsilon }n^2}{|x|}\),

  2. \(|\overline{E}_{-n}'(x)|\leq \frac{M_{r\varepsilon }n^2}{|x|}\),

where \(\overline{E}_{n}(x)=\sum _{m=1}^{n}{E}_{m}(x)\).

Proof â–¼
(i) Under conditions of this Lemma and Lemma 7 we have

\begin{eqnarray*} |\overline{E}_{n} ’(x)|& \leq & \sum _{m=1}^{n}|{E}_{m} ’(x)|\leq \frac{M_{r\varepsilon }}{|x|}\sum _{m=1}^{n}m\\ & =& \frac{M_{r\varepsilon }}{|x|}\cdot \frac{n(n+1)}{2}\leq \frac{M_{r\varepsilon }n^2}{|x|}, \end{eqnarray*}

for \(0{\lt}\varepsilon \leq |x|\leq \pi \).

(ii) Similarly we have obtained

\begin{eqnarray*} |\overline{E}_{-n} ’(x)|\leq \sum _{m=1}^{n}|{E}_{-m} ’(x)|\leq \frac{M_{r\varepsilon }}{|x|}\sum _{m=1}^{n}m\leq \frac{M_{r\varepsilon }n^2}{|x|}, \end{eqnarray*}

for \(0{\lt}\varepsilon \leq |x|\leq \pi \).

Proof â–¼

3 Main Results

The following theorem presents the main result.

Theorem 10

Let \((c_{k})\) belongs to the class \(\mathbb {K}^{2}\). Then

\begin{equation} \label{eq31} \lim _{n\to \infty }g_{n}^{c}(x)=f(x),\quad \text{exists for}\quad |x|\in (0,\pi ], \end{equation}
14

\begin{equation} \label{eq32} f\in L^{1}(0,\pi ]\quad \text{and}\quad \| g_{n}^{c}-f\| _{L^{1}}\to 0,\, \, \text{as}\, \, n\to \infty , \end{equation}
15

and

\begin{equation} \label{eq33} \| S_{n}(f)-f\| _{L^1}\to 0,\, \, \text{as}\, \, n\to \infty . \end{equation}
16

Proof â–¼
Firstly, we will show that \(f(x)\) exists on \((0,\pi ]\). Indeed, it is clear that we can write

\[ S_{n}(x)=\sum _{k=-n}^{n}c_{k}e^{ikx}=c_{0}+\sum _{k=1}^{n}\left(\frac{c_{k}}{k}ke^{ikx} +\frac{c_{-k}}{k}ke^{-ikx}\right) \]

Applying twice the Abel’s transformation we obtain

\begin{eqnarray} \label{eq34} S_{n}(x)\! \! \! & =& \! \! \! c_{0}-i\left[ \sum _{k=1}^{n-2}\bigtriangleup ^{2}\left(\frac{c_{k}}{k}\right)\overline{E}’_{k}(x)+\bigtriangleup \left(\frac{c_{n-1}}{n-1}\right)\overline{E}’_{n}(x)\right] -i\frac{c_{n}}{n}{E}’_{n}(x)\nonumber \\ & & +i\left[ \sum _{k=1}^{n-2}\bigtriangleup ^{2}\left(\frac{c_{-k}}{k}\right)\overline{E}’_{-k}(x)+\bigtriangleup \left(\frac{c_{-(n-1)}}{n-1}\right)\overline{E}’_{-n}(x)\right] +i\frac{c_{-n}}{n}{E}’_{-n}(x)\nonumber . \end{eqnarray}

Based on Lemmas 7 and 9 we clearly have

\begin{eqnarray} \label{eq35} \! \! \! \! & & \! \! \! \! |S_{n}(x)|\leq \nonumber \\ \! \! \! \! & \leq & \! \! \! \! |c_{0}|+\sum _{k=1}^{n-2}\Bigg[\left|\bigtriangleup ^{2}\left(\frac{c_{k}}{k}\right)\right| |\overline{E}’_{k}(x)|+\left|\bigtriangleup ^{2}\left(\frac{c_{-k}}{k}\right)\right| |\overline{E}’_{-k}(x)|\Bigg] \nonumber \\ & & +\Bigg[\left| \bigtriangleup \left( \frac{c_{n-1}}{n-1}\right) \right| |\overline{E}’_{n}(x)|+\left| \bigtriangleup \left( \frac{c_{-(n-1)}}{n-1} \right) \right| |\overline{E}’_{-n}(x)| \Bigg] \nonumber \\ & & +\frac{|c_{n}|}{n}|{E}’_{n}(x)|+\frac{|c_{-n}|}{n}|{E}’_{-n}(x)|\nonumber \\ \! \! \! \! & \leq & \! \! \! \! |c_{0}|+\frac{M_{r\varepsilon }}{|x|}\Bigg\{ \sum _{k=1}^{n-2}k^2 \Bigg[\left|\bigtriangleup ^{2}\left(\frac{c_{k}}{k}\right)\right| +\left|\bigtriangleup ^{2}\left(\frac{c_{-k}}{k}\right)\right| \Bigg] \nonumber \\ & & +n^2\Bigg[\left| \bigtriangleup \left( \frac{c_{n-1}}{n-1}\right) \right| +\left| \bigtriangleup \left( \frac{c_{-(n-1)}}{n-1} \right) \right| \Bigg] +|c_{n}|+|c_{-n}| \Bigg\} \nonumber \\ \! \! \! \! & \leq & \! \! \! \! |c_{0}|+\frac{2M_{r\varepsilon }}{|x|}\Bigg\{ \sum _{k=1}^{n-2}A_{k}+2\sum _{k=n-1}^{\infty }k^2 \Bigg[\left|\bigtriangleup ^{2}\left(\frac{c_{k}}{k}\right)\right| +\left|\bigtriangleup ^{2}\left(\frac{c_{-k}}{k}\right)\right| \Bigg] +2\overline{M} \Bigg\} \nonumber \\ \! \! \! \! & \leq & \! \! \! \! |c_{0}|+\frac{2M_{r\varepsilon }}{|x|}\Bigg\{ 5\sum _{k=1}^{\infty }k^2A_{k}+2\overline{M} \Bigg\} {\lt}+\infty ,\nonumber \end{eqnarray}

since \((c_{k})\in \mathbb {K}^2\), where \(\overline{M}\) is a positive constant.

Subsequently, \(\lim _{n \to \infty }S_{n}(x)=\lim _{n\to \infty }g_{n}^{c}(x)=f(x)\) exists, because of the boundedness of the functions \(\frac{e^{inx}}{\sin x}\), \(\frac{E_{n}(x)}{\sin x}\), \(\frac{E_{-n}(x)}{\sin x}\) on \((0,\pi ]\), and thus (14) holds true.

Now we are going to prove (15). Indeed, for \(x\not=0\) we have

\begin{eqnarray*} f(x)-g_{n}^{c}(x)\! \! \! \! & =& \! \! \! \! \sum _{k=n+1}^{\infty }\left(\frac{c_{k}}{k}ke^{ikx} +\frac{c_{-k}}{k}ke^{-ikx}\right) \\ & & -\frac{i}{2\sin x}[ c_{n}e^{i(n+1)x}- c_{-n}e^{-i(n+1)x}+c_{n+1}e^{inx}\\ & & - c_{-(n+1)}e^{-inx}+(c_n-c_{n+2})E_{n}(x)+(c_{-(n+2)}-c_{-n})E_{-n}(x)]. \end{eqnarray*}

Again, applying twice the Abel’s transformation to the above equality we obtain

\begin{eqnarray*} \! \! \! \! & & \! \! \! \! f(x)-g_{n}^{c}(x)= \nonumber \\ \! \! \! \! & =& \! \! \! \! -i\lim _{p\to \infty }\Bigg\{ \sum _{k=n+1}^{p-2}\left[\bigtriangleup ^{2}\left(\frac{c_{k}}{k}\right)\overline{E}’_{k}(x)-\bigtriangleup ^{2}\left(\frac{c_{-k}}{k}\right)\overline{E}’_{-k}(x)\right] \\ & & +\bigtriangleup \left(\frac{c_{p-1}}{p-1}\right)\overline{E}’_{p-1}(x) -\bigtriangleup \left(\frac{c_{-(p-1)}}{p-1}\right)\overline{E}’_{-(p-1)}(x) +\frac{c_p}{p}{E}’_{p}(x) \\ & & -\frac{c_{-p}}{p}{E}’_{-p}(x)-\bigtriangleup \left(\frac{c_{n}}{n}\right)\overline{E}’_{n}(x) +\bigtriangleup \left(\frac{c_{-n}}{n}\right)\overline{E}’_{-n}(x)- \frac{c_{n+1}}{n+1}{E}’_{n}(x)\\ & & +\frac{c_{-(n+1)}}{n+1}{E}’_{-n}(x)\Bigg\} -\frac{i}{2\sin x}[ c_{n}e^{i(n+1)x}- c_{-n}e^{-i(n+1)x}+c_{n+1}e^{inx}\\ & & -c_{-(n+1)}e^{-inx}+(c_n-c_{n+2})E_{n}(x)+(c_{-(n+2)}-c_{-n})E_{-n}(x)]\\ \! \! \! \! & =& \! \! \! \! -i\Bigg\{ \sum _{k=n+1}^{\infty }\left[\bigtriangleup ^{2}\left(\frac{c_{k}}{k}\right)\overline{E}’_{k}(x)-\bigtriangleup ^{2}\left(\frac{c_{-k}}{k}\right)\overline{E}’_{-k}(x)\right] -\bigtriangleup \left(\frac{c_{n}}{n}\right)\overline{E}’_{n}(x)\\ & & +\bigtriangleup \left(\frac{c_{-n}}{n}\right)\overline{E}’_{-n}(x)- \frac{c_{n+1}}{n+1}{E}’_{n}(x)+\frac{c_{-(n+1)}}{n+1}{E}’_{-n}(x)\Bigg\} \\ & & -\frac{i}{2\sin x}[ c_{n}e^{i(n+1)x}- c_{-n}e^{-i(n+1)x}+c_{n+1}e^{inx}-c_{-(n+1)}e^{-inx}\\ & & +(c_n-c_{n+2})E_{n}(x)+(c_{-(n+2)}-c_{-n})E_{-n}(x)]. \end{eqnarray*}

Hence, using Lemmas 7 and 9 we get

\begin{eqnarray*} \! \! \! \! & & \! \! \! \! |f(x)-g_{n}^{c}(x)| \leq \\ \! \! \! \! & \leq & \! \! \! \! \sum _{k=n+1}^{\infty }\left[\left|\bigtriangleup ^{2}\left(\frac{c_{k}}{k}\right)\right| |\overline{E}’_{k}(x)|+\left|\bigtriangleup ^{2}\left(\frac{c_{-k}}{k}\right)\right| |\overline{E}’_{-k}(x)|\right] \\ & & +\left|\bigtriangleup \left(\frac{c_{n}}{n}\right)\right| |\overline{E}’_{n}(x)| +\left|\bigtriangleup \left(\frac{c_{-n}}{n}\right)\right| |\overline{E}’_{-n}(x)|+\left| \frac{c_{n+1}}{n+1}\right| |{E}’_{n}(x)|\\ & & +\left|\frac{c_{-(n+1)}}{n+1}\right| |{E}’_{-n}(x)|\Bigg\} +\frac{1}{2|\sin x|}[ |c_{n}|+|c_{-n}|+|c_{n+1}|\\ & & +|c_{-(n+1)}|+(|c_n|+|c_{n+2}|)|E_{n}(x)|+(|c_{-(n+2)}|+|c_{-n}|)|E_{-n}(x)|]\\ \! \! \! \! & \leq & \! \! \! \! \frac{M_{r\varepsilon }}{|x|}\Bigg\{ \sum _{k=n+1}^{\infty }k^2\left[\left|\bigtriangleup ^{2}\left(\frac{c_{k}}{k}\right)\right| +\left|\bigtriangleup ^{2}\left(\frac{c_{-k}}{k}\right)\right| \right] \\ & & +n^2\left[\left|\bigtriangleup \left(\frac{c_{n}}{n}\right)\right| +\left|\bigtriangleup \left(\frac{c_{-n}}{n}\right)\right|\right] +(n+1)\left(\left| c_{n+1}\right| +\left| c_{-(n+1)}\right|\right) \Bigg\} \\ & & +\frac{|c_{n}|+|c_{-n}|+|c_{n+1}|+|c_{-(n+1)}|}{2|\sin x|} \end{eqnarray*}
\begin{eqnarray*} & & +(|c_n|+|c_{n+2}|)\left|\frac{E_{n}(x)}{2\sin x}\right| +(|c_{-(n+2)}|+|c_{-n}|)\left|\frac{E_{-n}(x)}{2\sin x}\right|\\ \! \! \! \! & \leq & \! \! \! \! \frac{M_{r\varepsilon }}{|x|}\Bigg[2\sum _{k=n}^{\infty }A_{k} +(n+1)\left(\left| c_{n+1}\right| +\left| c_{-(n+1)}\right|\right) \Bigg] \\ & & +\frac{|c_{n}|+|c_{-n}|+|c_{n+1}|+|c_{-(n+1)}|}{2|\sin x|} \end{eqnarray*}
\begin{eqnarray*} & & +(|c_n|+|c_{n+2}|)\left|\frac{E_{n}(x)}{2\sin x}\right| +(|c_{-(n+2)}|+|c_{-n}|)\left|\frac{E_{-n}(x)}{2\sin x}\right| . \end{eqnarray*}

Therefore, using Lemma 8 we obtain

\begin{eqnarray*} \| f-g_{n}^{c}\| _{L^{1}}\! \! \! \! & \leq & \! \! \! \! M_{r\varepsilon }\Bigg[2\sum _{k=n}^{\infty }A_{k}\int _{0}^{\pi }\frac{dx}{|x|} +(n+1)\left(\left| c_{n+1}\right| +\left| c_{-(n+1)}\right|\right) \int _{0}^{\pi }\frac{dx}{|x|} \Bigg] \\ & & +\frac{|c_{n}|+|c_{-n}|+|c_{n+1}|+|c_{-(n+1)}|}{2}\int _{0}^{\pi }\frac{dx}{|\sin x|}\\ & & +(|c_n|+|c_{n+2}|)\int _{0}^{\pi }\left|\frac{E_{n}(x)}{2\sin x}\right| dx \\ & & +(|c_{-(n+2)}|+|c_{-n}|)\int _{0}^{\pi }\left|\frac{E_{-n}(x)}{2\sin x}\right| dx\\ \! \! \! \! & \leq & \! \! \! \! M_{r\varepsilon }\Bigg[2\sum _{k=n}^{\infty }A_{k}\log k +(n+1)\left(\left| c_{n+1}\right| +\left| c_{-(n+1)}\right|\right) o\left( \log n\right) \Bigg] \\ & & +\frac{|c_{n}|+|c_{-n}|+|c_{n+1}|+|c_{-(n+1)}|}{2}o\left( \log n\right) \\ & & +(|c_n|+|c_{n+2}|)o(n) +(|c_{-(n+2)}|+|c_{-n}|)o(n). \end{eqnarray*}

Now we note that

\[ \sum _{k=n}^{\infty }A_{k}\log k\leq \sum _{k=n}^{\infty }k^2A_{k}=o(1), \]

and

\begin{eqnarray*} (n+1)\left| c_{\pm (n+1)}\right| \log n\! \! \! \! & \leq & \! \! \! \! (n+1)^3\left| \frac{c_{\pm (n+1)}}{n+1}\right|\\ \! \! \! \! & \leq & \! \! \! \! (n+1)^3\sum _{k=n+1}^{\infty }\left| \bigtriangleup \left(\frac{c_{\pm k}}{k}\right)\right|\\ \! \! \! \! & \leq & \! \! \! \! (n+1)^3\sum _{k=n+1}^{\infty }\sum _{j=k}^{\infty }\left| \bigtriangleup ^2\left(\frac{c_{\pm j}}{j}\right)\right|\\ \! \! \! \! & =& \! \! \! \! (n+1)^3\sum _{j=n+1}^{\infty }(j-n)\left| \bigtriangleup ^2\left(\frac{c_{\pm j}}{j}\right)\right| \end{eqnarray*}
\begin{eqnarray*} \! \! \! \! & \leq & \! \! \! \! \sum _{j=n+1}^{\infty }j^4\left| \bigtriangleup ^2\left(\frac{c_{\pm j}}{j}\right)\right|\\ \! \! \! \! & \leq & \! \! \! \! \sum _{j=n+1}^{\infty }j^4\frac{A_{j}}{j^2}=\sum _{j=n+1}^{\infty }j^2A_{j}=o(1) \end{eqnarray*}

as \(n\to \infty \).

Subsequently, we get

\[ \| f-g_{n}^{c}\| _{L^{1}}=o(1)\quad \text{as}\quad n\to \infty . \]

Using the latest equality and the fact that \(g_{n}^{c}(x)\) is a polynomial it follows that \(f\in L^1(0,\pi ]\).

Finally, we will prove (16). Namely, using some facts used above we have

\begin{eqnarray*} & & \int _{0}^{\pi }|f(x)-{S}_{n}(x)|dx\leq \\ & & \leq \int _{0}^{\pi }|f(x)-g^c_{n}(x)|dx+\int _{0}^{\pi }|g^c_{n} (x)-{S}_{n}(x)|dx\\ & & \leq \int _{0}^{\pi }|f(x)-g^c_{n}(x)|dx+\int _{0}^{ \pi }\bigg|\frac{i}{2\sin x}[ c_{n}e^{i(n+1)x}- c_{-n}e^{-i(n+1)x}+c_{n+1}e^{inx}\\ & & \hspace{.5cm}-c_{-(n+1)}e^{-inx}+(c_n-c_{n+2})E_{n}(x)+(c_{-(n+2)}-c_{-n})E_{-n}(x)]\bigg| dx\\ & & \leq \int _{0}^{\pi }|f(x)-g^c_{n}(x)|dx+[ |c_{n}|+| c_{-n}|+|c_{n+1}|+|c_{-(n+1)}|]\int _{0}^{ \pi }\frac{dx}{2|\sin x|}\\ & & \hspace{.5cm}+(|c_n|+|c_{n+2}|)\int _{0}^{ \pi }\bigg|\frac{E_{n}(x)}{2\sin x}\bigg| dx +(|c_{-(n+2)}|+|c_{-n}|)\int _{0}^{ \pi }\bigg|\frac{E_{-n}(x)}{2\sin x}\bigg| dx \\ & & \leq \int _{0}^{\pi }|f(x)-g^c_{n}(x)|dx+[ |c_{n}|+| c_{-n}|+|c_{n+1}|+|c_{-(n+1)}|]o\left( \log n\right)\\ & & \hspace{.5cm}+[|c_n|+|c_{n+2}|+|c_{-(n+2)}|+|c_{-n}|]o\left( n\right) \\ & & =o(1)+o(1)+o(1)=o(1), \, \, n\to \infty . \end{eqnarray*}

The proof of theorem is completed.

Proof â–¼

Acknowledgements

The author would like to thank the anonymous referee for her/his remarks which improved the presentation of this paper. Also, many thanks goes for my ex-supervisor, Professor Naim L. Braha, for his advices.

Bibliography

1

B. Ram, Convergence of certain cosine sums in the metric space \(L\), Proc. Amer. Math. Soc., 66 (1977), pp. 258–260. \includegraphics[scale=0.1]{ext-link.png}

2

B. Ram, Integrability of Ress-Stanojević sums, Acta Sci. Math., 42 (1980), pp. 153–155.

3

C. S. Rees and Č. V. Stanojević, Necessary and sufficient condition for intergrability of certain cosine sums, J. Math. Anal. Appl., 43 (1973), pp. 579–586. \includegraphics[scale=0.1]{ext-link.png}

4

L. Leindler, Additions to the Telyakovskiı̆ class \(S\), J. Ineq. Pure and Appl. Math., 4 (2003) Art. 35.

5

N. K. Bary, A treatise on trigonometric series (London: Pergamon Press) Vol.II, 1964.

6

S. A. Telyakovskiı̆ , On a sufficient condition of Sidon for integrability of trigonometric series, Math. Zametki, (in Russian) 14 (1973), pp. 317–328.

7

S. S. Bhatia and B. Ram, The extensions of the F. Móricz’s theorems, Proc. Amer. Math. Soc. 124 (1996), no. 6, pp. 1821–904. \includegraphics[scale=0.1]{ext-link.png}

8

S. Kumari and B. Ram, \(L^{1}\)-convergence of certain trigonometric sums, Indian J. Pure Appl. Math. 20 (9), 908–914, 1989.

9

J. Kaur and S. S. Bhatia, Convergence of modified complex trigonometric sums in the metric space \(L\), Lobachevskii J. Math., 31 (2010), No. 3, pp. 290–294. \includegraphics[scale=0.1]{ext-link.png}

10

J. Kaur and S. S. Bhatia, A class of \(L^1\) convergence of new modified cosine sum, Southeast Asian Bull. Math., 36 (6) (2012), pp. 831–836.

11

K. Kaur, S. S. Bhatia and B. Ram, Integrability and \(L^1\) convergence of modified sine sums, Georgian. Math. J. 11 (1) (2004), pp. 99–104. \includegraphics[scale=0.1]{ext-link.png}

12

T. Kano, Coefficients of some trigonometric series, J. Fac. Sci. Shinshu Univ., 3 (1968), pp. 153–162.

13

Xh. Z. Krasniqi, \(L^1\)-convergence of a set of modified sums, An. Univ. Oradea Fasc. Mat., Tomus XXII (2015), No. 1, pp. 41–46.

14

Xh. Z. Krasniqi, Some new modified cosine sums and \(L^1\)-convergence of cosine trigonometric series, Arch. Math. (Brno), Tomus 49 (2013), pp. 43–50.

15

Xh. Z. Krasniqi, On \(L^{1}\)-convergence of Rees-Stanojevic’s sums with coefficients from the class \(K\), Matematiche (Catania), Vol. LXV–Fasc. II (2010), pp. 25–32. \includegraphics[scale=0.1]{ext-link.png}

16

Ž. Tomovski, On a theorem of S. S. Bhatia and B. Ram, Math. Commun, 8, 187–193, 2003.