Professional Documents
Culture Documents
Spline Based Series For Sine and Arbitrarily Accurate Bounds For Sine, Cosine and Sine Integral
Spline Based Series For Sine and Arbitrarily Accurate Bounds For Sine, Cosine and Sine Integral
Roy M. Howard
School of Electrical Engineering, Computing and Mathematical Sciences, Curtin University, GPO Box
U1987, Perth, 6845, Australia.
email: r.howard@curtin.edu.au
17 June 2020
Abstract
Based on two point spline approximations of arbitrary order, a series of functions that define lower bounds for
sin x and sin x x , over the interval 0 2 , with increasingly low relative errors and smaller relative errors
than published results, are defined. Second, fourth and eighth order approximations have, respectively, maximum
–4 –8 – 18
relative errors over the interval 0 2 of 3.31 10 , 2.48 10 and 2.02 10 . New series for the sine
function, which have significantly better convergence that a Taylor series over the interval 0 2 , are pro-
posed. Applications include functions that are upper bounds for the sine function, upper and lower bounds for the
cosine function and lower bounds for the sine integral function. These bounded functions can be made arbitrarily
accurate.
Keywords: Jordan’s inequality, trigonometric inequalities, spline approximation, sine integral function.
1 Introduction
Research to establish upper and lower bounds for sin x x over the interval 0 2 has a long history (e.g.
Sándor & Bencze 2005, Mortici 2011) and Qi et al. 2009 provides a detailed review. Fundamental results include
Jordan’s inequality
2 sin x
--- --------------- 1 x 0 2 (1)
x
the Cusa-Huygens inequality
sin x 2 + cos x
--------------- ------------------------- 0 x --- (2)
x 3 2
and the Redheffer inequality (Redheffer 1969):
2 2
–x sin x
----------------- --------------- x 0 (3)
+x
2 2 x
Similar, and related inequalities, which are indicative of published bounds, are detailed in Table 1.1. The graphs
of the relative errors in the bounds defined by the first, second, fourth, fifth, eighth and tenth inequalities stated in
this table are shown in Figure 1. The relative error of the other bounds specified in Table 1.1 lie within the limits
of the bounds shown in Figure 1. The relative errors associated with these bounds typical have maximum values,
–4
over the interval 0 2 , in the range of 10 to 0.2 . Some of the bound are ‘sharp’, i.e. exact, at the points
zero and 2 . The majority of the bounds have relative errors that are monotonically increasing and have maxi-
mum values at 2 .
Applications of such bounds include defining bounds to the Riemann zeta function 3 , e.g. Luo et al., 2003,
and bounds on some trigonometric ratios, e.g. Wu, 2004.
lower 5
re x 1 lower 2
lower 1 upper 5
0.100
upper 1,2
lower 10
0.010 lower 8
upper 8
0.001 upper 5
lower 4
10-4
upper 10
10-5
upper 4
10-6
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 x
Figure 1 Graph of the magnitude of the relative errors in the upper and
lower bounds to sin x x as defined by the first, second, fourth, fifth,
eighth and tenth order functions defined in Table 1.1.
Naturally, it is of interest to determine bounds with lower relative errors. Zhu (2008a, Theorem 5) proposed the
following general form for bounds of sin x x , 0 x 2 :
n n n 2 2 n+1
2 2 k 2 2 n+1 sin x 2 2 k 2k – 4x
k – 4x + n + 1 – 4x ---------------
x k – 4x + 1 – ak -----------------------------------
2n + 2
k=0 k=0 k=0
(4)
2k – 1 1 2 1
k = --------------- k – 1 – -------------------------------- k – 2 0 = --- 1 = ----- n 0 1 2
2k
2
16 k – 1 k
2
3
and, as indicated in Figure 2, the relative error in these bounds can be made arbitrarily small by utilizing increas-
ingly high orders. This result was generalized in Zhu 2008b. These bounds build on the results of Li, 2006 (Theo-
rem 2.1) and an alternative, but equivalent form, was published by Nui et al. (2008). Nui et al. 2010 provides a
generalization and showed that the equation forms specified by Nui and Zhu are equivalent (Proposition 3). These
bounds, for orders zero to two, are:
2 2
2 – 4x sin x 2 1 – 2 2 2
--- + -------------------- --------------- --- + ------------------- – 4x (5)
3 x
2
2 2 2 2 2
2 – 4x 12 – 2 2 2 sin x 2 – 4x 1–3 2 2 2
--- + -------------------- + ----------------- – 4x --------------- --- + -------------------- + ------------------- – 4x (6)
3
16
5 x
3
4
2 2 2 2
2 – 4x 12 – 2 2 2 10 – 2 2 3 sin x
--- + -------------------- + ----------------- – 4x + ----------------- – 4x ---------------
3
16
5
16
7 x
(7)
2 2 2 2
12 –
– 4x - ----------------- 3 12 –
2 2 2 3
--2- + -------------------
2 1 2
+ – 4x + ----- 1 – --- – ----------------- – 4x
3
16
5
6 16
re x lower 0
0.010 upper 0
lower 1
0.001
upper 1
10-4
lower 2
-5
10
upper 2
10-6
10-7
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 x
This paper utilizes the two point spline based function approximation, detailed in Howard 2019, to specify a series
of upper and lower bounds for sin x and sin x x , over the interval 0 2 , with increasingly low relative
errors. Analysis associated with the bounds leads to new series for the sine function which have significantly bet-
ter convergence, over the interval 0 2 , than a Taylor series and the published series proposed by Zhu 2008a
and Nui 2008. Applications include functions that define upper and lower bounds for the cosine function and
lower bounds for the sine integral function. These are can be made arbitrarily accurate.
Section 2 details miscellaneous results that underpin subsequent analysis. Section 3 details a spline based series
for the sine function and in Section 4 it is shown that these series are lower bounds, of increasing accuracy. Sec-
tion 5 details applications including new series for the sine function, upper bounds for the sine function, bounds
for the cosine function and a lower bound for the sine integral function. Conclusions are detailed in section 6.
1.1 Notation
For a function f defined over the interval , an approximating function f A has a relative error, at a point x 1 ,
defined according to
fA x1
re x 1 = 1 – --------------- (8)
f x1
The relative error bound for the approximating function f A , over the interval , is defined according to
re B = max re x 1 : x 1 (9)
k
x = d f x is used.
k
The notation f
k
dt
Mathematica has been used to facilitate analysis and to obtain numerical results. In general, relative error bound
results for approximations have been obtained by sampling the appropriate interval at 1000 equally spaced
points.
2 Miscellaneous Results
The following results underpin later analysis and discussion:
3 5 7 k2 k
x x x –1 x
sin x = x – ----- + ----- – ----- + ----------------------------- + k 1 3 5 7 (10)
3! 5! 7! k!
A Taylor series yields alternating lower and upper bounds over the interval 0 2 as the order is increased
(e.g. Qi, 2009, Section 1.6). The bounds are increasingly accurate as can be seen in the results shown, along with
other results, in Figure 3. The relative error bounds, over the interval 0 2 , for various orders of Taylor series
approximations, are detailed in Table 2.1. The rate of convergence of the Taylor series serves as a reference.
Relative error
Order of approx. bound
1: upper bound 0.571
3: lower bound –2
7.52 10
5: upper bound –3
4.52 10
Relative error
Order of approx. bound
7: lower bound –4
1.57 10
9: upper bound –6
3.54 10
13: lower bound – 10
6.63 10
17: lower bound – 14
4.35 10
33: upper bound – 34
7.07 10
U L L
f k x = 2f k x – f k – 1 x x k 2 3 (11)
L L
provided k – 1 x 2 k x , x where
L L
k x = f x – fk x x (12)
Proof
U L L
Consider f k x = 2f k x – f k – 1 x and the error
U U L L
k x = f k x – f x = 2f k x – f k – 1 x – f x
(13)
L L L L
= 2f k x – 2f x + f x – f k – 1 x = – 2 k x + k – 1 x
L L U
When k – 1 x 2 k x it follows that f k x 0 as required.
f L --- – y cos y f U --- – y y 0 --- (14)
2 2 2
Proof
As sin 2 – x = cos x and sin x = cos 2 – x , for x 0 2 , it follows that
f L x sin x f U x f L x cos --- – x f U x x 0 ---
2 2
(15)
f L --- – y cos y f U --- – y y = --- – x x = --- – y y 0 ---
2 2 2 2 2
y 0 x = sin x = f 0 x + 0 x x 0 ---
2
(16)
y 1 x 1 = sin --- x 1 = f 1 x 1 + 1 x 1 x 1 0 1
2
2
y 0 x = y 1 --- x y 1 x 1 = y 0 --- x 1
2
(17)
2
0 x = 1 --- x 1 x 1 = 0 --- x 1
2
The relative errors in the approximations sin x f 0 x 1 and sin x 1 2 f 1 x 1 , respectively, are:
f0 x f1 x1
re 0 x = 1 – --------------- re 1 x 1 = 1 – ---------------------------- (18)
sin x sin x 1 2
re 0 x = re 1 2x re 1 x 1 = re 0 x 1 2 (19)
Proof
Consider, for example:
f 0 x 1 2 f1 x1
re 0 x 1 2 = 1 – ---------------------------- = 1 – ---------------------------- = re 1 x 1 (20)
sin x 1 2 sin x 1 2
2.4.1 Note
It is clear that establishing the relative error for an approximation to sin x 1 2 for x 1 0 1 is equivalent to
establishing the relative error for the scaled approximation to sin x for x 0 2 .
n+1 n k n–k i
– x x – k n + i ! x –
f n x = ---------------------------- ------------------- f ----------------- ------------------- +
–
n+1 k! i! n! – i
k=0 i=0
(21)
n+1 n k k n–k i
x – –1 – x k – x
n + i ! ------------------
---------------------------- -------------------------------- f ----------------- - x
–
n+1 k! i! n! – i
k=0 i=0
This general result can be utilized to define a sequence of functions that approximate, with increasing accuracy,
the sine function:
n n–k
2x n + 1 1 k k -----------------
n + i ! 2x i
f n x = 1 – ------
---- sin ------ x
k! 2 i! n!
------ +
k=0 i=0
n k n–k (22)
2x n + 1 –1 k k + 1 2x k n + i ! 2x i
------
------------- --- sin -------------------- 1 – ------
k! 2 2 -----------------
i! n!
1 – ------
k=0 i=0
n 0 1 2
The approximations, of order zero to fourth, are:
2x
f 0 x = ------ (23)
12 2 16 3
f 1 x = x + ------ 1 – --- x – ------ 1 – --- x (24)
2 3
3 4
2 2 2
80 3 3 240 4 4 192 5
f 2 x = x + ------ 1 – ------ – ------ x – --------- 1 – ------ – ------ x + --------- 1 – --- – ------ x (25)
3 10 80
4 15 60
5 4 48
3 2 3 2 3
x 560 2 4 2688 15 5
f 3 x = x – ----- + --------- 1 – ------ – ------ + --------- x – ------------ 1 – --------- – ------ + --------- x +
6
4 7 56 420
5 56 48 672
(26)
2 3 2 3
4480 9 13 6 2560 7
------------ 1 – ------ – ------------ + --------- x – ------------ 1 – --- – ------ + --------- x
6 35 560 840
7 4 40 960
3 2 3 4
x 4032 5 5 5
f 4 x = x – ----- + ------------ 1 – ------ – ------ + ------------ + --------------- x –
6
5 18 48 2016 48384
2 3 4 2 3 4
--------------- 1 – 4
26880 ------ – 11 - --------------
------------ + -------- - 6 69120 7 53 7
+ x + --------------- 1 – ------ – ------------ + --------- + --------------- x – (27)
6 15 480 504 40320
7 27 2160 576 34560
2 3 4 2 3 4
--------------- 1 – 16
80640 --------- – 13 - --------------
------------ + -------- - 8 35840 3 9
+ x + --------------- 1 – --- – --------- + --------- + --------------- x
8 63 504 630 30240
9 4 112 672 26880
Proof
These results arise from direct application of Equation 21 for the case of f x = sin x , the interval 0 2 ,
= 0 , = 2 and the result
k k
f x = sin x + ------ k 0 1 2 (28)
2
3.1 Results
The graphs of the relative errors in the first to fourth order spline based approximations, defined in Theorem 3.1,
are shown in Figure 3 and the associated relative error bounds are detailed in Table 3.1. The error functions for the
first to fourth order approximations are shown in Figure 4. These results, when compared with Taylor series
approximations and the approximations defined by Zhu 2008a (specified in Equation 4 and whose relative errors
are shown in Figure 2), demonstrate better convergence. The approximations are exact at the points 0 2 .
The errors in the approximations, for orders one to four, are shown in Figure 4 and are positive which suggests the
approximations can serve as lower bounds for the sine function. The proof of this is detailed below.
Relative error
Order of bound: Spline
approx. based.
0 0.363
1 –2
1.63 10
2 –4
3.31 10
3 –6
3.62 10
4 –8
2.48 10
6 – 13
3.91 10
8 – 18
2.02 10
16 – 43
9.19 10
32 – 100
2.19 10
Taylor 1 Taylor 3
fk x
1 – --------------- 0.010 order 1
sin x Taylor 5
0.001 order 2
-4
Taylor 7
10
10-7
order 4
10-8
0.025 4
10 3 x
0.020 100 2 x
0.015 6
10 4 x
0.010
1 x
0.005
0.000
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 x
Figure 4 Graph of the errors, k x = sin x – f k x , in the first
to fourth order spline approximations to sin x .
1 x 1 = sin --- x 1 – f 1 --- x 1 x 1 0 1 (29)
2 2
which has the property of being positive on the interval 0 1 and being such that 1 0 = 1 1 = 0 :
2 2 3 3 4 3 5 2 6 2 7 3
1 t = c2 t 1 – t + c3 t 1 – t + c4 t 1 – t + c5 t 1 – t + c6 t 1 – t + c7 t 1 – t +
k p 2 k = 2 5 6 9 10 13 14 17 18
= ck t 1 – t p =
3 k = 3 4 7 8 11 12 15 16 19 20
k=2 (30)
k + 1 2
5 –1
--- + ----------------------------------
k 2 2
= ck t 1 – t t 0 1
k=2
The series converges for t 0 1 and the coefficients can be defined iteratively according to
c0 = –1 c 1 = – 2 + --- = 2c 0 + --- c 2 = – 3 = 2c 1 – c 0 (31)
2 2
3 3 3
3 7
c 3 = – 4 + ------ – ------ = c 2 + c 1 – c 0 – ---------- c 4 = – 9 + ------ – ------ = 3c 3 – c 2 (32)
2 48 3
2 3! 2 16
and
c k = 3c k – 1 – c k – 2 k 4 8 12 16
k
c k = 2c k – 1 – c k – 3 + ---------- k 5 9 13 17
k
2 k!
(33)
c k = 2c k – 1 – 3c k – 2 + c k – 3 k 6 10 14 18
k
c k = 3c k – 2 – 4c k – 3 – c k – 4 + c k – 5 – ---------- k 7 11 15 19
k
2 k!
By construction the coefficients are positive and decrease according to 0 c k + 1 c k , k 2 , and it is the case
that 1 x 1 0 for x 1 0 1 .
Proof
The proof is detailed in Appendix 1.
4.1.1 Coefficients
The first few coefficients are: c 2 = 0.14159 , c 3 = 0.0664249 , c 4 = 0.057682 , c 5 = 0.053464 ,
–4 –4
c 6 = 3.06823 10 , and c 7 = 1.49925 10 .
4.1.2 Implication
As 1 x 1 0 , 0 x 1 1 , it follows that f 1 x 1 2 is a lower bound for sin x 1 2 , x 1 0 1 . The scaling
relationships specified in Lemma 3 then imply that f 1 x is a lower bound for sin x , x 0 2 .
2 x 1 = sin --- x 1 – f 2 --- x 1 x 1 0 1 (34)
2 2
which has the property of being positive on the interval 0 1 and being such that 2 0 = 2 1 = 0 :
3 3 4 4 5 4 6 3 7 3 8 4
2 t = c3 t 1 – t + c4 t 1 – t + c5 t 1 – t + c6 t 1 – t + c7 t 1 – t + c8 t 1 – t +
k p 3 k = 3 6 7 10 11 14 15
= ck t 1 – t p =
4 k = 4 5 8 9 12 13 16 17
k=3 (35)
k2
7 –1
--- + -----------------------
k 2 2
= ck t 1 – t
k=3
The coefficients can be defined iteratively according to
2 2
c0 = –1 c 1 = – 2 + --- = 2c 0 + --- c 2 = – 3 + + ----- = 2c 1 – c 0 + ---------- (36)
2 2 8 2
2 2!
2 3 3 2 3
c 3 = – 10 + 3 + ----- – ------ = c 2 + 4c 1 – c 0 – ---------- c 4 = – 15 + 5 + ----- – ------ = 3c 3 – 2c 2 – 4c 1 – c 0 (37)
8 48 3
2 3! 8 16
2 3 5 5
25 3
c 5 = – 36 + --------- + ----- – --------- + ---------- = 3c 4 – c 2 – 3c 1 + ---------- (38)
2 4 16 2 5 5! 5
2 5!
Proof
The proof is detailed in Appendix 2.
4.2.1 Coefficients
–3
The first few coefficients are: c 3 = 0.0125144 , c 4 = 0.00377153 , c 5 = 0.003325 , c 6 = 3.18534 10 and
–5
c 7 = 1.02411 10 .
4.2.2 Implication
As 2 x 1 0 , 0 x 1 1 , it follows that f 2 x 1 2 is a lower bound for sin x 1 2 , x 1 0 1 . The scaling
relationships specified in Lemma 3 then imply that f 2 x is a lower bound for sin x , x 0 2 .
5 Applications
3 4 5
1699x 11x 7x
f 2 x x – ----------------- + ------------ + ------------ (40)
10000 2000 1250
–4
yields a relative error bound of 6.59 10 . In comparison, a fifth order Taylor series approximation (Table 2.1)
– 3.
has a relative error bound of 4.52 10 .
–6
To achieve a relative error bound of 10 , a fourth order approximation, as specified by Equation 27, is required
–8
and the achieved relative error bound is 2.48 10 . The following fourth order approximation, with the resolu-
–7
tion defined by the given coefficient accuracy, yields a relative error bound of 4.31 10 :
3
x –3 5 –6 6 –4 7
f 4 x x – ----- + 8.33165 10 x + 5.17 10 x – 2.0463 10 x +
6 (41)
–6 8 –6 9
3.55 10 x + 1.89 10 x
–6
A ninth order Taylor series approximation (Table 2.1) yields a relative error bound of 3.54 10 .
– 10 – 16
For relative error bounds of 10 and 10 , spline approximations of orders six and eight are required. The
required approximations are:
3 5
x x –4 7 – 10 8 –6 9
f 6 x x – ----- + --------- – 1.98412876 10 x + 7.78 10 x + 2.754282 10 x +
6 120 (42)
– 9 10 – 8 11 – 10 12 – 10 13
1.485 10 x – 2.5944 10 x + 3.06 10 x + 1.11 10 x
3 5 7
x x x –6 9
f 8 x x – ----- + --------- – ------------ + 2.755731916334 10 x +
6 120 5040
– 14 10 – 8 11 – 13 12 – 10 13 (43)
3.43891 10 x – 2.50521948 10 x + 1.26122 10 x + 1.604729071 10 x +
– 14 14 – 13 15 – 15 16 – 15 17
7.21759 10 x – 7.936185 10 x + 7.0722 10 x + 1.956 10 x
– 11 – 17
The achieved relative error bounds, respectively, are 4.39 10 and 5.66 10 . Higher precision in the
– 13 – 18
coefficients leads to relative error bounds, respectively, of 3.91 10 and 2.02 10 . For comparison, a
thirteenth and a seventeenth order Taylor series approximation have, respectively, relative error bounds of
– 10 – 14
6.63 10 and 4.35 10 .
12 16 2 sin x
0 1 + ------ 1 – --- x – ------ 1 – --- x --------------- x 0 2 (44)
2 3
3 4 x
2 2 2
80 3 2 240 4 3 192 4 sin x
0 1 + ------ 1 – ------ – ------ x – --------- 1 – ------ – ------ x + --------- 1 – --- – ------ x --------------- x 0 2 (45)
3 10 80
4 15 60
5 4 48 x
and similar results hold for higher order spline approximations. There is equality at the points 0 2 . The maxi-
mum relative errors associated with these lower bounds are detailed in Table 3.1.
Proof
This result follows from Theorem 4.1 and Theorem 4.2 and, similarly, for higher order spline approximations.
Theorem 5.2 Series Approximation for Sin Based on First Order Spline Approximation
The error function associated with a first order spline approximation yields the following series for the sine func-
tion that is valid for the interval 0 2 :
2x 2x 2x 2x k 2x p 2 k = 1 2 5 6 9 10 13 14 17
sin x = ------ + ------ 1 – ------ +
c k ------ 1 – ------
p =
3 k = 3 4 7 8 11 12 15 16 19
(46)
k=1
2x 2x 2x 2x 2x 2 2x 2 2x 2
sin x = ------ + ------ 1 – ------ + 2 – 1 + --- ------ 1 – ------ + – 3 ------ 1 – ------ +
4
(47)
3 3 3 4
3 2x 2x 3 7 2x 2x 3
– 4 + ------ – ------ ------ 1 – ------ + – 9 + ------ – ------ ------ 1 – ------ +
2 48 2 16
Proof
The proof is detailed in Appendix 3.
n
2x 2x 2x 2x k 2x p 2 k = 1 2 5 6 9 10
sin x ------ + ------ 1 – ------ +
c k ------ 1 – ------
p =
3 k = 3 4 7 8 11 12
(48)
k=1
The graphs of the relative errors associated with the approximations to sin x , which are defined by the first to
ninth terms in the series approximations, are shown in Figure 5. The associated relative error bounds are detailed
in Table 5.1.
Table 5.1 Relative error bounds for the first and second
order series approximations to sin x .
re x
1
0.001
2
10-5
3
4
10-7
5 6 8 9
10-9
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 x
Figure 5 Graphs of the relative errors in the approximation to
sin x , as defined by Equation 48, for the cases of the first to
ninth terms.
Theorem 5.3 Series Approximation for Sin Based on Second Order Spline Approximation
The error function associated with a second order spline approximation yields the following series for the interval
0 2 :
2
2x 2 2x k 2x p 3 k = 2 3 6 7 10 11 14 15
sin x = 1 – ----- 1 – ------ +
8 c k ------ 1 – ------
p =
4 k = 0 1 4 5 8 9 12 13 16 17
(49)
k=0
where
2 2 2
3
c 0 = – 1 + ----- c 1 = – 4 + --- + ----- c 2 = – 10 + 2 + --------- (50)
8 2 4 8
and with the coefficients c 3 c 4 being defined in Theorem 4.2. The first few terms are
2 2 2
2x 2 2x 4 2x
------ 1 – 2x
4
sin x = 1 – ----- 1 – ------ + – 1 + ----- 1 – ------ + – 4 + --- + ----- ------ +
8 8 2 4
(51)
2 2 2 3 3
3 2x 2x 3 2x 2x 3
– 10 + 2 + --------- ------ 1 – ------ + – 10 + 3 + ----- – ------ ------ 1 – ------ +
8 8 48
Proof
The proof is detailed in Appendix 4.
2 n
2x 2 2x k 2x p 3 k = 2 3 6 7 10 11
sin x 1 – ----- 1 – ------ +
8 c k ------ 1 – ------
p =
4 k = 0 1 4 5 8 9
(52)
k=0
The graphs of the relative errors in the approximations, for the case of two to nine terms, are shown in Figure 6.
The associated relative error bounds are detailed in Table 5.1.
re x
10-4
2
-5
10
10-6
3
-7
10
4
10-8
5 6 7
10-9 9
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 x
Figure 6 Graphs of the relative errors in the approximations to
sin x as defined by Equation 52, for orders two to nine.
2 2 2
80 3 3 240 4 4 192 5
f 2 x = x + ------ 1 – ------ – ------ x – --------- 1 – ------ – ------ x + --------- 1 – --- – ------ x (53)
3 10 80
4 15 60
5 4 48
2x 2x 2x 2x 2x 2 2x 2 2x 2
f 2 x = ------ + ------ 1 – ------ + 2 – 1 + --- ------ 1 – ------ + – 3 ------ 1 – ------ (54)
4
2 2 2
2x 2 2x 4 2x 2x 4
f 2 x = 1 – ----- 1 – ------ + – 1 + ----- 1 – ------ + – 4 + --- + ----- ------ 1 – ------ +
8 8 2 4
(55)
2 2
3 2x 2x 3
– 10 + 2 + --------- ------ 1 – ------
8
–4 –3 –4
with respective relative error bounds over the interval 0 2 of 3.31 10 , 2.70 10 and 3.31 10 . In
general, the series specified by Theorem 3.1 have better convergence that the other two series and a simpler form.
2x 2x
1 ------ 2 2 ------ x 0 --- (56)
2
where 1 and 2 are defined by Equation 30 and Equation 35. When this is the case the following upper bound
can be defined:
U
f 2 x = 2f 2 x – f 1 x x 0 2 (57)
i.e.
2 2 2
12 176 13 2 480 4 3 384 4 sin x
1 – ------ 1 – --- x + --------- 1 – --------- – ------ x – --------- 1 – ------ – ------ x + --------- 1 – --- – ------ x ---------------
2 3
3 44 88
4 15 60
5 4 48 x (58)
x 0 2
The error in this bound, as defined by
U U
2 x = f 2 x – sin x x 0 2 (59)
U
is shown in Figure 7. Note that the error 2 x is close to the error in the lower bound, 1 x , associated with
f 1 x , which is shown in Figure 4.
Proof
The proof is detailed in Appendix 5.
0.025 4 U
10 4 x
0.020
0.015
U
2 x
0.010 U
100 3 x
0.005
0.000
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 x
U U U
Figure 7 Graphs of the errors in the functions f 2 f 3 and f 4 ,
which defined second, third and fourth order upper bounds to
sin x .
U
f 3 x = 2f 3 x – f 2 x x 0 2
(60)
U
f 4 x = 2f 4 x – f 3 x x 0 2
are valid upper bounds with f 2 f 3 and f 4 being defined by Equation 25 to Equation 27. The graphs of the errors
U U
in these bounds, denoted, respectively, 3 and 4 , are shown in Figure 7.
12 2 16 3
g 1 y = 1 – ------ 1 – --- y + ------ 1 – --- y (61)
2 6
3 4
2 2 2 2
y 80 3 3 240 7 4 192 5
g 2 y = 1 – ----- – ------ 1 – --- – --------- y + --------- 1 – ------ – ------ y – --------- 1 – --- – ------ y (62)
2 3 5 80
4 30 40
5 4 48
2 2 3 2 3
y 560 3 4 2688 13 5 5
g 3 y = 1 – ----- – --------- 1 – ------ – ------ + ------------ y + ------------ 1 – --------- – --------- + ------------ y –
2 4 14 28 1680
5 56 168 1344
(63)
2 3 2 3
4480-
-----------
17 3 6 2560 7
1 – --------- – --------- + ------------ y + ------------ 1 – --- – ------ + --------- y
6 70 112 1120
7 4 40 960
2 4 2 3 4
y y 4032 2 5 5 5
g 4 y = 1 – ----- + ------ – ------------ 1 – ------ – --------- + ------------ + --------------- y +
2 24
5 9 144 1008 48384
2 3 4 2 3 4
26880- 7 23 6 69120 13 7 11 7
-------------- 1 – ------ – ------ + --------------- + --------------- y – --------------- 1 – --------- – --------- + ------------ + --------------- y + (64)
6 30 32 20160 16128
7 54 240 8640 20736
2 3 4 2 3 4
80640 31 8 35840 3 9
--------------- 1 – --------- – ------ + --------- + --------------- y – --------------- 1 – --- – --------- + --------- + --------------- y
8 126 36 720 24192
9 4 112 672 26880
Proof
These results follow from the result stated in Lemma 2 of f L 2 – y cos y , y 0 2 and the spline
approximations for the sine function stated in Theorem 3.1. Together these imply g k y cos y where
g k y = f k 2 – y , k 1 2 3 4 . Alternatively, they can be directly derived from Equation 21 for the case
of f x = cos x .
5.7.1 Results
The relative error bounds in the approximations for the cosine function defined in Theorem 5.5 are identical to the
relative error bounds associated with the lower bounds for the sine function which are detailed in Table 3.1.
2 2 2 2
12 2 176 9 3 3 480 7 4 384 5
1 + ------ 1 – --- – ------ y – --------- 1 – ------ – --------- y + --------- 1 – ------ – ------ y – --------- 1 – --- – ------ y
2 6 12
3 44 88
4 30 40
5 4 48 (65)
cos y y 0 2
The error in this bound is
U U
2 cos y = g 2 y – cos y = 2g 2 y – g 1 y – cos y
(66)
U
= 2f 2 --- – y – f 1 --- – y – sin --- – y = 2 --- – y
2 2 2 2
U
and, thus, has the same bound as 2 whose graph is shown in Figure 7. Higher order upper bounds can be defined
in a similar manner.
x
sin
Si x = ---------------- d
(67)
0
The lower bound published by Lv 2017 is
3
2x + sin x ----------------------------------------------------------
x + 3x cos x – 3 sin x -
--------------------------- – Si x (68)
3 9
2
Integration of the spline series for sin x defined by Theorem 3.1 (after dividing by x ) leads to the following
approximations:
6 2 16 3
h 1 x = x + ----- 1 – --- x – --------- 1 – --- x (69)
2 3 3
3 4
2 2 2
80 3 3 60 4 4 192 5
h 2 x = x + --------- 1 – ------ – ------ x – ------ 1 – ------ – ------ x + --------- 1 – --- – ------ x (70)
3
2 10 80
4 15 60 5
5 4 48
3 2 3 2 3
x 140 2 4 2688 15 5
h 3 x = x – ------ + --------- 1 – ------ – ------ + --------- x – ------------ 1 – --------- – ------ + --------- x +
18 4 7 56 420 5
5 56 48 672
(71)
2 3 2 3
2240 9 13 - 6 2560
------------ 1 – ------ – ------------ + --------
7
x – ------------ 1 – --- – ------ + --------- x
3
6 35 560 840 7
7 4 40 960
3 2 3 4
x 4032 5 5 5
h 4 x = x – ------ + ------------ 1 – ------ – ------ + ------------ + --------------- x –
18 5 5 18 48 2016 48384
2 3 4 2 3 4
4480 4 11 - --------------
------------ 1 – ------ – ------------ + -------- - 6 69120 7 53 7
+ x + --------------- 1 – ------ – ------------ + --------- + --------------- x – (72)
6 15 480 504 40320 7
7 27 2160 576 34560
2 3 4 2 3 4
10080 16 13 8 35840 3 9
--------------- 1 – --------- – ------------ + --------- + --------------- x + --------------- 1 – --- – --------- + --------- + --------------- x
8 63 504 630 30240 9
9 4 112 672 26880
Proof
First to fourth order spline based approximations for the sine integral, and for the interval 0 2 , can be gener-
ated from the approximations detailed in Theorem 3.1 for the sine function according to
x
fk
hk x = ------------ d
k 0 1 (73)
0
The stated results then follow from the definitions for f k given in Theorem 3.1. As f k , k 0 1 are lower
bounds for the sine function, it then follows that the approximations for the sine integral function are also lower
bounds.
5.8.1 Results
Graphs of the relative error in the approximations defined in Theorem 5.6 are shown in Figure 8 and show the
improvement over the lower bound proposed by Lv 2017 for orders three and higher. Bounds on the relative error
are tabulated in Table 5.2 and clearly can be made arbitrarily small by increasing the order of approximation.
hk x
1 – ------------- 0.010 order 1
si x
0.001
order 2
10-4 Lv 2017
10-5
order 3
10-6
10-7
10-8 order 4
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 x
Figure 8 Graph of the relative error in spline based
approximations, of orders one to four, for the sine integral function
as well as the relative error in the lower bound specified by Lv
(2017) and stated in Equation 68.
5.8.2 Notes
Upper bound functions for the sine integral can be generated from the lower bounds and consistent with
Lemma 1.
6 Conclusion
Spline based approximations for sin x over the interval 0 2 have been to detailed and exhibit better con-
vergence that comparative Taylor series approximations. It was proved that these approximations are also lower
bounds for sin x and sin x x , over the interval 0 2 , with increasingly high accuracy and with lower rel-
ative error bounds than published lower bounded functions. For example, approximations of order two, four and
–4 –8 – 18
eight have relative error bounds, respectively, of 3.31 10 , 2.48 10 and 2.02 10 .
The analysis underpinning the proof that the approximations are lower bounds for the sine function led to new
series for sin x . These have better convergence that Taylor series for sin x over the interval 0 2 .
Applications were detailed and include, first, a sequence of upper bounded functions, with increasing accuracy,
for sin x over the interval 0 2 . Second, a sequence of upper and lower bounded functions for cos x with
accuracy consistent with the corresponding sine approximations. Finally, a sequence of convergent lower
bounded approximations for the sine integral function over the interval 0 2 .
Acknowledgement The support of Prof. A. Zoubir, SPG, Technische Universität Darmstadt, Darmstadt, Ger-
many, who hosted a visit where the writing of this paper was completed, is gratefully acknowledged.
1 t = sin --- t – f 1 --- t
2 2
3 5 1 7 9
2 3 1 1 (74)
= – 3 t + 2 – --- – ------ t + ----- --- t – ----- --- t + ----- --- t +
2 48 5! 2 7! 2 9! 2
2 3 5 7 9
= e2 t + e3 t + e5 t + e7 t + e9 t +
3
where e 2 = – 3 = 0.314159 and e 3 = 2 – 2 – 48 = – 0.21676 etc. The goal is to write 1 as a sum-
mation of positive terms only. To this end the following form is considered:
2 p2 3 p3 4 p4 n pn
1 t = c2 t 1 – t + c3 t 1 – t + c4 t 1 – t + + cn t 1 – t +
p2 2 p2 p3 p4
= c2 t 1 – p2 t + t – t + c3 t 1 – p3 t + t + c4 t 1 – p4 t + t +
2 3 4
2 (75)
p2
= c 2 t + – p 2 c 2 + c 3 t + c 2 – p 3 c 3 + c 4 t +
2 3 4
2
Equating Equation 74 and Equation 75 it immediately follows that c 2 = e 2 0 . For c 3 to be positive, the factor
p 2 needs to be chosen such that e 3 = – p 2 c 2 + c 3 and c 3 = e 3 + p 2 c 2 0 . This is guaranteed when
p 2 – e 3 c 2 . Choosing the smallest integer greater than zero such that this is the case yields
p2 = – e 3 e 2 + 1 and for the given values of e 2 and e 3 the required value is p 2 = 2 . It then follows that
3
c 3 = – 4 + 3 2 – 48 = 0.0664249 .
c3 e3 + p2 c2 e3 e3 –e –e3 –e
----- = ---------------------- = ----- + p 2 = ----- + -------3- + 1 = 1 – -------- – -------3- (76)
c2 e2 e2 e2 e2 e 2 e2
p3 2 p3 p4
1 t = c 2 t 1 – 2t + t + c 3 t 1 – p 3 t + t – t + c 4 t 1 – p 4 t + t
2 2 3 4
2
(77)
p3 2 p3 p4
= e 2 t + e 3 t + c 2 – p 3 c 3 + c 4 t + c 3 t t – t + c 4 t – p 4 t + t +
2 3 4 3 4 2
2
c4 –c2 + p3 c3 –c c c2
- = -------2- + p 3 = 1 +
----- = ------------------------- ----2- – ----- (78)
c3 c3 c3 c3 c3
which implies 0 c 4 c 3 1 .
Proceeding in this manner yields the stated form and proof of positive coefficients with decreasing values. The
coefficients are defined according to
3
3
c2 = – 3 + c 3 = – 4 + ------ – ------ (79)
2 48
3 3 5
7
c 4 = – 9 + ------ – ------ c 5 = – 15 + 6 – ----- + ------------ (80)
2 16 8 3840
3 5 3 5 7
7 5
c 6 = – 7 + 3 – ------ + ------------ c 7 = – 8 + ------ – --------- + ------------ – ------------------ (81)
12 1920 2 48 1280 645120
3 5 7 3 5 7 9
15 11 7 3
c 8 = – 17 + --------- – ------------ + ------------ – ------------------ c 9 = – 27 + 12 – --------- + --------- – ------------------ + --------------------------------- (82)
2 48 3840 215040 8 320 107520 185 794 560
3 5 7 9
c 10 = – 11 + 5 – ----- + --------- – ------------------ + ------------------------------ (83)
6 640 161280 92 897 280
3 5 7 9 11
11 3 7
c 11 = – 12 + --------- – --------- + ------------ – ------------------ + ------------------------------ – ------------------------------------------ (84)
2 16 3840 129024 61 931 520 81 749 606 400
3 5 7 9 11
23 19 11
c 12 = – 25 + --------- – ------------ + --------- – ------------------ + ------------------------------ – ------------------------------------------ (85)
2 48 256 645120 26 542 080 27 249 868 800
To find a general formula for the coefficients consider the definitions:
c0 = –1 c 1 = – 2 + --- = 2c 0 + --- (86)
2 2
It then can be shown that
3
c 2 = 2c 1 – c 0 c 3 = c 2 + c 1 – c 0 – ---------- (87)
3
2 3!
5
c 4 = 3c 3 – c 2 c 5 = 2c 4 – c 2 + ---------- (88)
5
2 5!
7
c 6 = 2c 5 – 3c 4 + c 3 c 7 = 3c 5 – 4c 4 – c 3 + c 2 – ---------- (89)
7
2 7!
The iteration formulas
c k = 3c k – 1 – c k – 2 k 4 8 12 16
k
c k = 2c k – 1 – c k – 3 + ---------- k 5 9 13 17
k
2 k!
(90)
c k = 2c k – 1 – 3c k – 2 + c k – 3 k 6 10 14 18
k
c k = 3c k – 2 – 4c k – 3 – c k – 4 + c k – 5 – ---------- k 7 11 15 19
k
2 k!
then follow.
2 2 3 3 4 3 5 2 6 2 7 3
1 t = c2 t 1 – t + c3 t 1 – t + c4 t 1 – t + c5 t 1 – t + c6 t 1 – t + c7 t 1 – t +
2 3 4 5 6 7
c2 t + c3 t + c4 t + c5 t + c6 t + c7 t + (91)
2 2 3 4 5 2 1
c 2 t 1 + t + t + t + t + t + = c 2 t ---------- t 0 1
1–t
The inequalities follow as c k 0 , 0 c k + 1 c k 1 for all k and the final form, consistent with the convergence
of a geometric series, is valid for t 0 1 . As 1 0 = 1 1 = 0 it follows that the series defining 1 t is
bounded above on the interval t 0 1 .
By utilizing a Taylor series expansion for sin t 2 , t 0 1 , and the definition for f 2 defined by Equation 25,
2 t = sin --- t – f 2 --- t
2 2
2 3 2 2 5 7 9
3 4 3 5 1 1 (92)
= – 10 + 3 + ----- – ------ t + 15 – 4 – ----- t + – 6 + ------ + ----- + ------------ t – ----- --- t + ----- --- t +
8 48 4 2 8 3840 7! 2 9! 2
3 4 5 7 9
= e3 t + e4 t + e5 t + e7 t + e9 t +
2 3 2
where e 3 = – 10 + 3 + 8 – 48 = 0.0125144 and e 4 = 15 – 4 – 4 = – 0.0337717 etc. The
approach is then to proceed in a manner consistent with the first order case detailed in Appendix 1 and the follow-
ing form is considered:
3 p3 4 p4 n pn
2 t = c3 t 1 – t + c4 t 1 – t + + cn t 1 – t +
p3 2 p3 p4 p5
= c3 t 1 – p3 t + t – t + c4 t 1 – p4 t + t + c5 t 1 – p5 t + t +
3 4 5
2 (93)
p3
= c 3 t + – p 3 c 3 + c 4 t + c 3 – p 4 c 4 + c 5 t +
3 4 5
2
Equating Equation 92 and Equation 93, it immediately follows that c 3 = e 3 0 . For the coefficient c 4 to be pos-
itive, the factor p 3 needs to be chosen such that e 4 = – p 3 c 3 + c 4 and c 4 = e 4 + p 3 c 3 0 . This is guaranteed
when p 3 – e 4 c 3 . Choosing the smallest integer greater than zero such this is the case yields
p 3 = – e 4 e 3 + 1 and for the defined values of e 3 and e 4 the required value is p 3 = 3 . It then follows that
2 3
c 4 = – 15 + 5 + 8 – 16 = 0.00377153 .
c4 e4 + p3 c3 e4 e4 –e –e4 –e
----- = ---------------------- = ----- + p 3 = ----- + -------4- + 1 = 1 – -------- – -------4- (94)
c3 e3 e3 e3 e3 e3 e3
which implies 0 c 4 c 3 1 . Proceeding in this manner yields the stated form and proof of positive coefficients
with decreasing values.
2 3 2 3
c 3 = – 10 + 3 + ----- – ------ c 4 = – 15 + 5 + ----- – ------ (95)
8 48 8 16
2 3 5 2 3 5
25 3 3 19
c 5 = – 36 + --------- + ----- – --------- + ------------ c 6 = – 64 + 23 + --------- – ------------ + --------- (96)
2 4 16 3840 8 48 960
2 3 5 7 2 3 5 7
5 7
c 7 = – 36 + 14 + ----- – --------- + --------- – ------------------ c 8 = – 45 + 18 + ----- – --------- + --------- – ------------------ (97)
8 16 640 645120 8 16 384 215040
2 3 5 7 9
81 49 5
c 9 = – 100 + --------- + ----- – ------------ + --------- – --------------- + --------------------------------- (98)
2 4 48 768 71680 185 794 560
2 3 5 7 9
3 85 23 19
c 10 = – 166 + 68 + --------- – ------------ + ------------ – ------------------ + ------------------------------ (99)
8 48 1920 645120 46 448 640
2 3 5 7 9 11
15 7
c 11 = – 78 + 33 + ----- – ------------ + --------- – --------------- + ------------------------------ – ------------------------------------------ (100)
8 16 960 43008 30 965 760 81 749 606 400
2 3 5 7 9 11
55 3
c 12 = – 91 + 39 + ----- – ------------ + --------- – --------------- + ------------------------------ – ------------------------------------------ (101)
8 48 320 30720 18 579 456 27 249 868 800
To find a general formula for the coefficients consider the definitions:
2 2
c0 = –1 c 1 = – 2 + --- = 2c 0 + --- c 2 = – 3 + + ----- = 2c 1 – c 0 + ----- (102)
2 2 8 8
It is then the case that
2 3 3
c 3 = – 10 + 3 + ----- – ------ = c 2 + 4c 1 – c 0 – ---------- (103)
8 48 3
2 3!
and it can then be readily shown that
5
c 4 = 3c 3 – 2c 2 – 4c 1 – c 0 c 5 = 3c 4 – c 2 – 3c 1 + ---------- (104)
5
2 5!
7
c 6 = 4c 5 – 6c 4 + c 3 c 7 = 2c 6 – 2c 5 – 2c 4 + c 3 – ---------- (105)
7
2 7!
9
c 8 = 3c 7 – 3c 6 + 4c 5 – c 4 c 9 = 4c 8 – 3c 7 + c 6 – c 5 + ---------- (106)
9
2 9!
The iterative formulas then follow:
2x
sin x = f 1 x + 1 ------ x 0 --- (108)
2
where f 1 is defined by Equation 24 and the convergent series for 1 is defined in Theorem 4.1. As f 1 can be
written in the form
2x 2x 2x 2x 2x 2
f 1 x = ------ + ------ 1 – ------ + c 1 ------ 1 – ------ c 1 = 2 – 1 + --- (109)
4
the required result follows, namely
2x 2x 2x 2x k 2x p 2 k = 1 2 5 6 9
sin x = ------ + ------ 1 – ------ +
c k ------ 1 – ------
p =
3 k = 3 4 7 8
(110)
k=1
2x
sin x = f 2 x + 2 ------ x 0 --- (111)
2
where f 2 is defined by Equation 25 and the convergent series for 2 is defined in Theorem 4.2, As f 2 can be
written in the form
2 2 2
2x 2 2x 4 2x
------ 1 – 2x
4
f 2 x = 1 – ----- 1 – ------ + – 1 + ----- 1 – ------ + – 4 + --- + ----- ------ +
8 8 2 4
(112)
2 2
3 2x 2x 3
– 10 + 2 + --------- ------ 1 – ------
8
the required result follows, namely:
2
2x 2 2x k 2x p
sin x = 1 – ----- 1 – ------ +
8 c k ------ 1 – ------
k=0
2 2 2
80 3 3 240 4 4 192 5
2f 2 x – f 1 x = 2 x + ------ 1 – ------ – ------ x – --------- 1 – ------ – ------ x + --------- 1 – --- – ------ x –
3 10 80
4 15 60
5 4 48
12 2 16 3
x + ------ 1 – --- x – ------ 1 – --- x (114)
2 3
3 4
2 2 2
12 2 176 13 3 480 4 4 384 5
= x – ------ 1 – --- x + --------- 1 – --------- – ------ x + --------- 1 – ------ – ------ x + --------- 1 – --- – ------ x
2 3
3 44 88
4 15 60
5 4 48
Second, utilizing the transformed functions associated with the interval 0 1 rather than the interval 0 2 ,
the requirement is for 1 t 2 2 t , t 0 1 , i.e. for
2 2 3 3 4 3 5 2 6 2 7 3
c2 t 1 – t + c3 t 1 – t + c4 t 1 – t + c5 t 1 – t + c6 t 1 – t + c7 t 1 – t +
(115)
3 3 4 4 5 4 6 3 7 3 8 4
2 d3 t 1 – t + d4 t 1 – t + d5 t 1 – t + d6 t 1 – t + d7 t 1 – t + d8 t 1 – t +
where the positive coefficients c k and d k are defined, respectively, in Theorem 4.1 and Theorem 4.2. Consider
k p1 k+1 p1 + 1
ck t 1 – t – 2d k + 1 t 1 – t p 1 2 3 t 0 1 k 2 3 4
k p1 k p1 k p1
= c k – 2d k + 1 t 1 – t t 1 – t = c k – 2d k + 1 t 1 – t + 2d k + 1 1 – t 1 – t t 1 – t (116)
k p1
c k – 2d k + 1 t 1 – t
7 References
Bercu, G., 2016, ‘Padé approximant related to remarkable inequalities involving trigonometric functions’, Journal
of Inequalities and Applications, 2016:99, doi: 10.1186/s13660-016-1044-x.
Bhayo, B. A., Klén, R. & Sándor, J., 2017, ‘New trigonometric and hyperbolic inequalities’, Miskolc
Mathematical Notes, vol. 18, no. 1, 2017, pp. 125-137, doi: 10.18514/MMN.2017.1560.
Chen., C. P. & Sándor, J., 2015, ‘Sharp inequalities for trigonometric and hyperbolic functions’, Journal of
Mathematical Inequalities, vol. 9, no. 1, pp. 203-217, doi: 10.7153/jmi-09-19.
Hau, Y., 2016, ‘Sharp Wilker and Huygens type inequalities for trigonometric and hyperbolic functions’,
Hacettepe Journal of Mathematics and Statistics, vol. 45, no. 3, pp. 731-741, doi: 10.15672/
HJMS.20164513099.
Howard, R. M., 2019, ‘Dual Taylor series, spline based function and integral approximation and applications’,
Mathematical and Computational Applications, vol. 24, article 35, doi:10.3390/mca24020035.
Kober, H., 1944, ‘Approximation by integral functions in the complex domain’, Transactions of the American
Mathematical Society, vol. 56, no. 1, pp. 7–31, eqn. 4.52.
Li, J. L., 2006, ‘An identity related to Jordan’s inequality’, International Journal of Mathematics and
Mathematical Sciences, 2006, Article ID: 76782, 6 pages, doi: 10.1155/IJMMS/2006/76782.
Luo, Q. M., Wei, Z. L. & Qi, F., 2003, ‘Lower and upper bounds of 3 ’, Advanced Studies in Contemporary
Mathematics, www.jangjeon.or.kr, vol. 6, no. 1, pp. 47-51.
Lv, H.L., Yang, Z.H., Luo, T.Q. & Zheng, S. Z., 2017, ‘Sharp inequalities for tangent function with applications’,
Journal of Inequalities and Applications, article ID: 2017:94, doi:10.1186/s13660-017-1372-5.
Mitrinović, D. S. & Adamović, D. D., 1965, ‘Sur une inégalité élémentaire où interviennent des fonctions
trigonométriques’, Univerziteta u Beogradu, Publikacije Elektrotehničkog fakulteta, Serija Matematika i
fizika, vol. 149, pp. 23–34.
Mortici, C., 2011, ‘The natural approach of Wilker-Cusa-Huygens inequalities’, Mathematical Inequalities and
Applications, vol. 14, no. 3, pp. 535-541.
Niu, D. W., Huo, Z. H., Cao, J. & Qi, F., 2008, ‘A general refinement of Jordan’s inequality and a refinement of L.
Yang’s inequality’, Integral Transforms and Special Functions, vol. 19, no. 3, pp. 157-164, doi: 10.1080/
10652460701635886.
Niu, D. W., Cao, J. & Qi, F., 2010, ‘Generalizations of Jordan’s inequality and concerned relations’, University
Politehnica of Bucharest, Scientific Bulletin, Series A, vol. 72, no. 3, pp. 85-98.
Qi, F., Cui, L. H. & Xu, S. L., 1999, ‘Some inequalities constructed by Tchebycheff's integral inequality’,
Mathematical Inequalities and Applications, vol. 2, no. 4, pp. 517-528, doi: 10.7153/mia-02-42.
Qi, F., Nui, D. W. & Guo, B. N., 2009, ‘Refinements, generalizations, and applications of Jordan's inequality and
related problems’, Journal of Inequalities and Applications, Article ID: 271923, 52 pages, doi:10.1155/2009/
271923.
Redheffer, R., 1969, ‘Problem 5642’, American Mathematical Monthly, vol. 76, no. 4, p. 422.