Next Article in Journal
Results on Second-Order Hankel Determinants for Convex Functions with Symmetric Points
Next Article in Special Issue
On Generalized Bivariate (p,q)-Bernoulli–Fibonacci Polynomials and Generalized Bivariate (p,q)-Bernoulli–Lucas Polynomials
Previous Article in Journal
On the Ideal Convergent Sequences in Fuzzy Normed Space
Previous Article in Special Issue
Bounds for Extreme Zeros of Classical Orthogonal Polynomials Related to Birth and Death Processes
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

On Some Bounds for the Gamma Function

Mathematics Department, Faculty of Science, King Abdulaziz University, P.O. Box 80203, Jeddah 21589, Saudi Arabia
*
Author to whom correspondence should be addressed.
Symmetry 2023, 15(4), 937; https://doi.org/10.3390/sym15040937
Submission received: 29 March 2023 / Revised: 15 April 2023 / Accepted: 18 April 2023 / Published: 19 April 2023

Abstract

:
Both theoretical and applied mathematics depend heavily on inequalities, which are rich in symmetries. In numerous studies, estimations of various functions based on the characteristics of their symmetry have been provided through inequalities. In this paper, we study the monotonicity of certain functions that involve Gamma functions. We were able to obtain some of the bounds of Γ ( v ) that are more accurate than some recently published inequalities.

1. Introduction

Mathematicians have made considerable efforts to develop more precise estimates of n ! and its natural extension, Gamma function. Scottish mathematician James Stirling (1692–1770) introduced the following formula:
n ! 2 π n n / e n , n
which is the most widely used and well-known approximation formula for handling large factorials and it bears his name [1,2,3]. Additionally, Stirling’s series [4]
Γ ( v + 1 ) 2 π v v / e v e p = 1 B 2 p 2 p ( 2 p 1 ) v 2 p 1 , v
is a generalization of formula (1), where B p denotes Bernoulli numbers. French scientist Pierre-Simon Laplace (1749–1827) [4] presented
Γ ( 1 + v ) 2 π v v / e v 1 + 1 12 v + 1 288 v 2 139 51,840 v 3 , v .
In 1917, Burnside [5] provided a more accurate formula than (1) with
Γ ( v + 1 ) 2 π v 2 v + 1 2 e v + 1 / 2 , v .
Indian mathematician Srinivasa Ramanujan (1887–1920) [6] presented the asymptotic expansion
Γ ( v + 1 ) π v e v 8 v 3 + 4 v 2 + v + 1 / 30 6 , v
and the following inequality of Gamma function between symmetric bounds
π v e v 8 v 3 + 4 v 2 + v + 1 / 100 6 < Γ ( v + 1 ) < π v e v 8 v 3 + 4 v 2 + v + 1 / 30 6 , v 0
which, according to the book The Lost Notebook and Other Unpublished Papers, are conjectures based on some mathematical calculations (see also [7,8,9,10,11]). In 2001, Karatsuba [9] presented
Γ ( v + 1 ) π v / e v 8 v 3 + 4 v 2 + v + 1 / 30 + θ 1 ( v ) 6 , v
where
θ 1 ( v ) = 11 240 v + 79 3360 v 2 + 3539 201,600 v 3 + ,
which proves Ramanujan’s formula (5). In 2011, Mortici [12] presented
Γ ( v + 1 ) π v / e v 8 v 3 + 4 v 2 + v + 1 / 30 6 e θ 2 ( v ) , v
where
θ 2 ( v ) = 11 11,520 v 4 + 13 3440 v 5 + 1 691,200 v 6 + ,
which improves Ramanujan formula (5) and is faster than formula (7).
In 2002, in web post, Robert H. Windschitl [13] (see also [14]) presented the important formula
Γ ( v + 1 ) = 2 π v v / e v v sinh v 1 v / 2 1 + O v 5 , v
which relates the Gamma function and the hyperbolic sine function. He advised using the approximation 2 π v v / e v v sinh v 1 + 1 810 v 6 v / 2 to calculate the values of the Gamma function on calculators with limited program or register memory since it is accurate to more than eight decimal places for v > 8 .
In 2009, Alzer [15] presented the following double inequality with a symmetrical bounds structure:
v sinh v 1 v / 2 1 + A 1 v 5 < Γ ( v + 1 ) 2 π v v / e v < v sinh v 1 v / 2 1 + A 2 v 5 , v > 0
with the best possible constants A 1 = 0 and A 2 = 1 1620 . Numerical calculations show that the lower bound in inequality (10) is superior to that of its counterpart in inequality (6) for v 2.07 . Additionally, the upper bound in inequality (10) is superior to that of its counterpart in inequality (6) for v 0.992 . In 2010, Nemes [16] presented
Γ ( v + 1 ) = 2 π v v / e v 1 + 10 120 v 2 1 v 1 + O v 5 , v
which is considerably easier than (9) and has exactly the same number of exact digits. Formulas (9) and (11) are more accurate than Ramanujan’s formula. In 2014, Lu, Song and Ma [17] deduced that there exists an n, such that for every v > n , the double inequality
v sinh 1 v + 1 810 v 7 67 42,525 v 9 v / 2 < Γ ( v + 1 ) 2 π v v / e v < v sinh 1 v + 1 810 v 7 v / 2
holds. Additionally, they provided some numerical comparisons to show how much better their approximations were than others such as Nemes’ formula (11). In 2022, Mahmoud and Almuashi [18] presented the new asymptotic formulas
Γ ( v + 1 ) 2 π v ( v / e ) v v 2 + 7 60 v 2 1 20 v / 2 exp r = 1 μ r v r , v
and
Γ ( v + 1 ) 2 π v ( v / e ) v v 2 + 7 60 v 2 1 20 v / 2 r = 0 λ r v r , v
where
μ r = B r + 1 r ( r + 1 ) 1 2 ρ r + 1 , ρ r = χ r 1 r j = 1 r 1 j ρ j χ r j , χ 0 = 1 , χ 2 r = 10 3 ( 20 ) r , χ 2 r 1 = 0 , λ 0 = 1 , λ r = 1 r j = 1 r j λ r j μ j , r = 1 , 2 , 3 , .
Both the two formulas (9) and (13) have the same rate of convergence, but the second is simpler.
For more details about asymptotic formulas and bounds of Γ ( v ) , please see [17,19,20,21,22,23] and the references therein.
In the rest of this paper, and motivated by formula (13), we will prove the following double symmetric inequality:
2 π v ( v / e ) v v 2 + 7 60 v 2 1 20 v / 2 1 + λ v 5 < Γ ( v + 1 ) < 2 π v ( v / e ) v v 2 + 7 60 v 2 1 20 v / 2 1 + μ v 5 , v 1
with λ = 0 and the best possible constant μ = 461 907,200 . Additionally, we will present comparisons between this inequality and the inequalities (6) and (10) presented by Ramanujan and Alzer, respectively, to clarify the superiority of our new results.

2. Main Results

Now, we will present new bounds of the Gamma function depending on the asymptotic formulas (12) and (13).
Theorem 1.
The function
P 1 ( v ) = Γ ( v + 1 ) 2 π v v e v v 2 + 7 60 v 2 1 20 v / 2
is strictly decreasing for v 1 . Furthermore,
Γ ( v + 1 ) > 2 π v ( v / e ) v v 2 + 7 60 v 2 1 20 v / 2 , v 1 .
Proof. 
The function T 1 ( v ) = log P 1 ( v ) log P 1 ( v + 1 ) satisfies T 1 ( v ) = F 1 ( v ) F 2 ( v ) , where
F 1 ( v + 1 ) = 4.4256 × 10 11 v 12 + 7.96608 × 10 12 v 11 + 6.48621 × 10 13 v 10 + 3.15729 × 10 14 v 9 + 1.02268 × 10 15 v 8 + 2.32054 × 10 15 v 7 + 3.77909 × 10 15 v 6 + 4.44603 × 10 15 v 5 + 3.74551 × 10 15 v 4 + 2.19984 × 10 15 v 3 + 8.53104 × 10 14 v 2 + 1.95514 × 10 14 v + 1.99325 × 10 13
and
F 2 ( v ) = 2 v 2 ( v + 1 ) 2 20 v 2 1 2 20 v 2 + 40 v + 19 2 60 v 2 + 7 2 60 v 2 + 120 v + 67 2 .
Then, T 1 ( v ) is a convex function for v 1 , and hence T 1 ( v ) is an increasing function for v 1 . Using the asymptotic expansion (12), we have
ln P 1 ( v ) 461 907,200 v 5 5197 9,072,000 v 7 + 1,436,249 1,710,720,000 v 9 26,863,154,077 14,010,796,800,000 v 11 + , v
and
P 1 ( v ) P 1 ( v ) 461 181,440 v 6 + 5197 1,296,000 v 8 1,436,249 190,080,000 v 10 + 26,863,154,077 1,273,708,800,000 v 12 + , v .
Then,
lim v T 1 ( v ) = 0
and T 1 ( v ) is negative for v 1 . So,
P 1 ( v ) P 1 ( v ) P 1 ( v + 1 ) P 1 ( v + 1 ) < 0 , v 1
and hence
P 1 ( v ) P 1 ( v ) < P 1 ( v + 1 ) P 1 ( v + 1 ) < P 1 ( v + 2 ) P 1 ( v + 2 ) < < lim n P 1 ( v + n ) P 1 ( v + n ) , v 1 .
However,
lim n P 1 ( v + n ) P 1 ( v + n ) = 0 .
Therefore, P 1 ( v ) < 0 or P 1 ( v ) is decreasing function for v 1 with lim v P 1 ( v ) = 1 , where we use the asymptotic expansion (13) to obtain
P 1 ( v ) 1 + 461 907,200 v 5 5197 9,072,000 v 7 + 1,436,249 1,710,720,000 v 9 + , v .
Then, P 1 ( v ) > 1 for v 1 or
Γ ( v + 1 ) > 2 π v ( v / e ) v v 2 + 7 60 v 2 1 20 v / 2 , v 1 .
Theorem 2.
The function
P 2 ( v ) = Γ ( v + 1 ) 2 π v v e v v 2 + 7 60 v 2 1 20 v / 2 461 907,200 v 5 + 1
is strictly increasing for v 1 . Furthermore,
Γ ( v + 1 ) < 2 π v ( v / e ) v v 2 + 7 60 v 2 1 20 v / 2 1 + μ v 5 , v 1
with the best possible constant μ = 461 907,200 .
Proof. 
The function T 2 ( v ) = ln P 2 ( v ) ln P 2 ( v + 1 ) satisfies T 2 ( v ) = F 3 ( v ) F 4 ( v ) , where
F 3 ( v + 1 ) = 8.11049 × 10 35 v 30 3.64972 × 10 37 v 29 7.91358 × 10 38 v 28 1.10102 × 10 40 v 27 1.10437 × 10 41 v 26 8.50657 × 10 41 v 25 5.23342 × 10 42 v 24 2.64083 × 10 43 v 23 1.11381 × 10 44 v 22 3.98112 × 10 44 v 21 1.21847 × 10 45 v 20 3.21822 × 10 45 v 19 7.37762 × 10 45 v 18 1.47414 × 10 46 v 17 2.5746 × 10 46 v 16 3.93687 × 10 46 v 15 5.27342 × 10 46 v 14 6.18444 × 10 46 v 13 6.3396 × 10 46 v 12 5.66425 × 10 46 v 11 4.39262 × 10 46 v 10 2.93963 × 10 46 v 9 1.68451 × 10 46 v 8 8.18045 × 10 45 v 7 3.32038 × 10 45 v 6 1.10538 × 10 45 v 5 2.93909 × 10 44 v 4 6.00088 × 10 43 v 3 8.83256 × 10 42 v 2 8.34133 × 10 41 v 3.79528 × 10 40
and
F 4 ( v ) = 2 v 2 907,200 v 5 + 4,536,000 v 4 + 9,072,000 v 3 + 9,072,000 v 2 + 4,536,000 v + 907,661 2 ( v + 1 ) 2 20 v 2 1 2 20 v 2 + 40 v + 19 2 60 v 2 + 7 2 60 v 2 + 120 v + 67 2 907,200 v 5 + 461 2 .
Then T 2 ( v ) is a concave function for v 1 and hence T 2 ( v ) is a decreasing function for v 1 . Using the asymptotic expansion (12), we have
ln P 2 ( v ) 5197 9,072,000 v 7 + 1,436,249 1,710,720,000 v 9 + 212,521 1,646,023,680,000 v 10 + , v
and
P 2 ( v ) P 2 ( v ) 5197 1,296,000 v 8 1,436,249 190,080,000 v 10 212,521 164,602,368,000 v 11 + , v .
Then
lim v T 2 ( v ) = 0
and T 2 ( v ) is positive for v 1 . So,
P 2 ( v ) P 2 ( v ) P 2 ( v + 1 ) P 2 ( v + 1 ) > 0 , v 1
and hence
P 2 ( v ) P 2 ( v ) > P 2 ( v + 1 ) P 2 ( v + 1 ) > P 2 ( v + 2 ) P 2 ( v + 2 ) > > lim n P 2 ( v + n ) P 2 ( v + n ) , v 1 .
However,
lim n P 2 ( v + n ) P 2 ( v + n ) = 0 .
Therefore, P 2 ( v ) > 0 or P 2 ( v ) is an increasing function for v 1 with lim v P 2 ( v ) = 1 , where we use the asymptotic expansion (13) to obtain
P 2 ( v ) 1 1 + 461 907,200 v 5 1 + 461 907,200 v 5 5197 9,072,000 v 7 + 1,436,249 1,710,720,000 v 9 + , v .
Then, P 2 ( v ) < 1 for v 1 or
Γ ( v + 1 ) < 2 π v ( v / e ) v v 2 + 7 60 v 2 1 20 v / 2 1 + 461 907,200 v 5 , v 1 .
This inequality is equivalent to T 3 ( v ) < μ for all v 1 , where
T 3 ( v ) = v 5 Γ ( v + 1 ) 2 π v ( v / e ) v v 2 + 7 60 v 2 1 20 v / 2 1 .
However, using the asymptotic expansion (13), we obtain
lim v T 3 ( v ) = 461 907,200 .
This implies that μ 461 907,200 , or the best possible value of μ is 461 907,200 . □

3. Comparison between Previous and New Results

Remark 1.
Using the expansion
v 2 + 7 60 v v 2 1 20 sinh v 1 = p = 2 ( p 1 ) ( 2 p + 7 ) 10 ( 2 p + 3 ) ! v 2 p ,
then we obtain
v 2 + 7 60 v 2 1 20 > v sinh v 1 , v > 1 20
and our new lower bound of inequality (15) is better than the lower bound of Alzer’s inequality (10) for v 1 .
Remark 2.
Consider the function
T 4 ( v ) = ln v 2 + 7 60 v 2 1 20 v / 2 461 907,200 v 5 + 1 ln v 4 + 31 294 v 2 v 4 3 49 v 2 + 11 5880 v / 2 1 1620 v 5 + 1 ,
then we have T 4 ( v ) = F 5 ( v ) F 6 ( v ) , where
F 5 ( v + 1 ) = 3.14926 × 10 33 v 32 + 1.00776 × 10 35 v 31 + 1.56197 × 10 36 v 30 + 1.56184 × 10 37 v 29 + 1.13218 × 10 38 v 28 + 6.339 × 10 38 v 27 + 2.85184 × 10 39 v 26 + 1.05892 × 10 40 v 25 + 3.30784 × 10 40 v 24 + 8.81677 × 10 40 v 23 + 2.02672 × 10 41 v 22 + 4.05079 × 10 41 v 21 + 7.08347 × 10 41 v 20 + 1.0888 × 10 42 v 19 + 1.47617 × 10 42 v 18 + 1.76937 × 10 42 v 17 + 1.87751 × 10 42 v 16 + 1.76448 × 10 42 v 15 + 1.46798 × 10 42 v 14 + 1.07968 × 10 42 v 13 + 7.00358 × 10 41 v 12 + 3.99296 × 10 41 v 11 + 1.99146 × 10 41 v 10 + 8.63451 × 10 40 v 9 + 3.22806 × 10 40 v 8 + 1.02952 × 10 40 v 7 + 2.76159 × 10 39 v 6 + 6.11223 × 10 38 v 5 + 1.08669 × 10 38 v 4 + 1.49173 × 10 37 v 3 + 1.48399 × 10 36 v 2 + 9.52017 × 10 34 v + 2.95688 × 10 33
and
F 6 ( v ) = v 1 20 v 2 2 60 v 2 + 7 2 294 v 2 + 31 2 5880 v 4 360 v 2 + 11 2 1620 v 5 + 1 2 907,200 v 5 + 461 2 .
Then T 4 ( v ) is concave for v 1 or T 4 ( v ) is decreasing for v 1 . However,
T 4 ( v ) 1793 42,336,000 v 8 793,529 96,808,320,000 v 10 11,231 18,289,152,000 v 11 + 1,328,563,181 1,280,774,073,600,000 v 12 + , v .
Then lim v T 4 ( v ) = 0 and T 4 ( v ) > 0 for v 1 or T 4 ( v ) is increasing for v 1 with lim v T 4 ( v ) = 0 , where
T 4 ( v ) 1793 296,352,000 v 7 + 793,529 871,274,880,000 v 9 + 11,231 182,891,520,000 v 10 + , v .
Therefore, T 4 ( v ) < 0 for v 1 or
v 2 + 7 60 v 2 1 20 v / 2 461 907,200 v 5 + 1 < v 4 + 31 294 v 2 v 4 3 49 v 2 + 11 5880 v / 2 1 1620 v 5 + 1 , v 1 .
Using the following expansion for v > 0
11 v 4 5880 3 v 2 49 + 1 sinh ( v ) 31 v 3 294 + v = p = 4 ( ( p 3 ) ( p 2 ) ( 44 p ( p + 4 ) + 245 ) ) v 2 p + 1 1470 ( 2 p + 1 ) ! > 0
we have
v sinh v 1 v / 2 1 1620 v 5 + 1 > v 4 + 31 294 v 2 v 4 3 49 v 2 + 11 5880 v / 2 1 1620 v 5 + 1 , v 1 .
Using inequalities (17) and (18), we conclude that our new upper bound of inequality (15) is better than the upper bound of Alzer’s inequality (10) for v 1 .
Remark 3.
Consider the function
T 5 ( v ) = 1 6 ln 8 v 3 + 4 v 2 + v + 1 100 ln 2 v v 2 + 7 60 v 2 1 20 v / 2 ,
then we have T 5 ( v ) = F 7 ( v ) 6 v 20 v 2 1 60 v 2 + 7 800 v 3 + 400 v 2 + 100 v + 1 , where
F 7 ( v + 3.4 ) = 3.2256 × 10 10 v 11 1.10615 × 10 12 v 10 1.70749 × 10 13 v 9 1.5627 × 10 14 v 8 9.39317 × 10 14 v 7 3.87682 × 10 15 v 6 1.11372 × 10 16 v 5 2.20314 × 10 16 v 4 2.88463 × 10 16 v 3 2.28577 × 10 16 v 2 8.809 × 10 15 v 6.29583 × 10 14 .
Then T 5 ( v ) is concave for v 3.4 or T 5 ( v ) is decreasing for v 3.4 . However,
T 5 ( v ) 30,847 103,680,000 v 8 + 1 1,280,000 v 7 + 539 207,360 v 6 23 4800 v 5 + 7 4800 v 4 + , v .
Then lim v T 5 ( v ) = 0 and T 5 ( v ) > 0 for v 3.4 , or T 5 ( v ) is increasing for v 3.4 with lim v T 5 ( v ) = 0 , where
T 5 ( v ) 529 61,440,000 v 8 + 30,847 725,760,000 v 7 1 7,680,000 v 6 539 1,036,800 v 5 + 23 19,200 v 4 7 14,400 v 3 + , v .
Therefore, T 5 ( v ) < 0 for v 3.4 or
8 v 3 + 4 v 2 + v + 1 100 6 < 2 v v 2 + 7 60 v 2 1 20 v / 2 , v 3.4 .
Then our new lower bound of inequality (15) is superior to that of its counterpart in inequality (6) for v 3.4 .
Remark 4.
Consider the function
T 6 ( v ) = 1 6 ln 8 v 3 + 4 v 2 + v + 1 30 ln 2 v v 2 + 7 60 v 2 1 20 v / 2 461 907,200 v 5 + 1
then we have T 6 ( v ) = F 8 ( v ) 2 v 2 1 20 v 2 2 60 v 2 + 7 2 240 v 3 + 120 v 2 + 30 v + 1 2 907,200 v 5 + 461 2 , where
F 8 ( v + 1.5 ) = 2.6073 × 10 21 v 20 + 7.68646 × 10 22 v 19 + 1.0737 × 10 24 v 18 + 9.44632 × 10 24 v 17 + 5.86836 × 10 25 v 16 + 2.73516 × 10 26 v 15 + 9.91887 × 10 26 v 14 + 2.86407 × 10 27 v 13 + 6.68251 × 10 27 v 12 + 1.27107 × 10 28 v 11 + 1.97924 × 10 28 v 10 + 2.52336 × 10 28 v 9 + 2.62358 × 10 28 v 8 + 2.20574 × 10 28 v 7 + 1.47838 × 10 28 v 6 + 7.72565 × 10 27 v 5 + 3.04037 × 10 27 v 4 + 8.51281 × 10 26 v 3 + 1.5292 × 10 26 v 2 + 1.39397 × 10 25 v + 2.20424 × 10 23 .
Then T 6 ( v ) is convex for v 1.5 or T 6 ( v ) is increasing for v 1.5 . However,
T 6 ( v ) 2629 10,368,000 v 8 + 1 115,200 v 7 + 13 2688 v 6 11 2880 v 5 + , v .
Then lim v T 6 ( v ) = 0 and T 6 ( v ) < 0 for v 1.5 or T 6 ( v ) is decreasing for v 1.5 with lim v T 6 ( v ) = 0 , where
T 6 ( v ) 121 22,118,400 v 8 + 2629 72,576,000 v 7 1 691,200 v 6 13 13,440 v 5 + 11 11,520 v 4 + , v .
Therefore, T 6 ( v ) > 0 for v 1.5 or
8 v 3 + 4 v 2 + v + 1 30 6 > 2 v v 2 + 7 60 v 2 1 20 v / 2 461 907,200 v 5 + 1 , v 1.5 .
Then our new upper bound of inequality (16) is superior to that of its counterpart in inequality (6) for v 1.5 .

4. Conclusions

The main conclusions of this paper are stated in Theorems (1) and (2). Concretely speaking, we studied the monotonicity of two functions involving the Gamma function to introduce the double inequality (14). We proved that our new inequality is better than Alzer’s double inequality (10) for v 1 . Additionally, our new lower (upper) bound is better than the lower (upper) bound of Ramanujan’s inequality (6) for v 3.4 ( v 1.5 ), respectively. Our results demonstrate that the approximation formula (12) had some advantages over Windschitl’s formula (9) in producing more precise inequalities for the Gamma function.

Author Contributions

Writing—original draft preparation, M.M., S.M.A. and S.A.; writing—review and editing, M.M., S.M.A. and S.A. All authors contributed equally to the writing of this paper. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding.

Data Availability Statement

Not applicable.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Batir, N. Very accurate approximations for the factorial function. J. Math. Inequal. 2010, 4, 335–344. [Google Scholar] [CrossRef]
  2. Gosper, R.W. Decision procedure for indefinite hypergeometric summation. Proc. Natl. Acad. Sci. USA 1978, 75, 40–42. [Google Scholar] [CrossRef] [PubMed]
  3. Mortici, C. On Gospers formula for the Gamma function. J. Math. Inequal. 2011, 5, 611–614. [Google Scholar] [CrossRef]
  4. Abramowitz, M.; Stegun, I.A. Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables, 9th ed.; Nation Bureau of Standards, Applied Mathematical Series; Dover Publications: New York, NY, USA; Washington, DC, USA, 1972; Volume 55. [Google Scholar]
  5. Burnside, W. A rapidly convergent series for logN! Messenger Math. 1917, 46, 157–159. [Google Scholar]
  6. Andrews, G.E.; Berndt, B.C. Ramanujan’s Lost Notebook: Part IV; Springer Science+ Business Media: New York, NY, USA, 2013. [Google Scholar]
  7. Chen, C.-P.; Elezović, N.; Vukšić, L. Asymptotic formulae associated with the Wallis power function and digamma function. J. Class. Anal. 2013, 2, 151–166. [Google Scholar] [CrossRef]
  8. Berndt, B.C.; Choi, Y.-S.; Kang, S.-Y. The problems submitted by Ramanujan. J. Indian Math. Soc., Contemp. Math. 1999, 236, 15–56. [Google Scholar]
  9. Karatsuba, E.A. On the asymptotic representation of the Euler Gamma function by Ramanujan. J. Comput. Appl. Math. 2001, 135, 225–240. [Google Scholar] [CrossRef]
  10. Mortici, C. On Ramanujan’s large argument formula for the Gamma function. Ramanujan J. 2011, 26, 185–192. [Google Scholar] [CrossRef]
  11. Ramanujan, S. The Lost Notebook and Other Unpublished Papers; Narosa Publ. H.-Springer: New Delhi, India; Berlin/Heidelberg, Germany, 1988. [Google Scholar]
  12. Mortici, C. Improved asymptotic formulas for the Gamma function. Comput. Math. Appl. 2011, 61, 3364–3369. [Google Scholar] [CrossRef]
  13. Programmable Calcualtors. Available online: http://www.rskey.org/CMS/the-library/11 (accessed on 20 April 2020).
  14. Smith, W.D. The Gamma Function Revisited. 2006. Available online: http://schule.bayernport.com/gamma/gamma05.pdf (accessed on 20 April 2020).
  15. Alzer, H. Sharp upper and lower bounds for the Gamma function. Proc. R. Soc. Edinb. 2009, 139A, 709–718. [Google Scholar] [CrossRef]
  16. Nemes, G. New asymptotic expansion for the Gamma function. Arch. Math. 2010, 95, 161–169. [Google Scholar] [CrossRef]
  17. Lu, D.; Song, L.; Ma, C. A generated approximation of the Gamma function related to Windschitl’s formula. J. Number Theory 2014, 140, 215–225. [Google Scholar] [CrossRef]
  18. Mahmoud, M.; Almuashi, H. On Some Asymptotic Expansions for the Gamma Function. Symmetry 2022, 14, 2459. [Google Scholar] [CrossRef]
  19. Chen, C.-P. Asymptotic expansions of the Gamma function related to Windschitl’s formula. Appl. Math. Comput. 2014, 245, 174–180. [Google Scholar] [CrossRef]
  20. Yang, Z.-H.; Tian, J.-F. An accurate approximation formula for Gamma function. J. Inequal. Appl. 2018, 2018, 56. [Google Scholar] [CrossRef] [PubMed]
  21. Yang, Z.-H.; Tian, J.-F. Two asymptotic expansions for Gamma function developed by Windschitl’s formula. Open Math. 2018, 16, 1048–1060. [Google Scholar] [CrossRef]
  22. Yang, Z.-H.; Tian, J.-F. A family of Windschitl type approximations for Gamma function. J. Math. Inequal. 2018, 12, 889–899. [Google Scholar] [CrossRef]
  23. Yang, Z.-H.; Tian, J.-F. Windschitl type approximation formulas for the Gamma function. J. Inequal. Appl. 2018, 2018, 272. [Google Scholar] [CrossRef] [PubMed]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Mahmoud, M.; Alsulami, S.M.; Almarashi, S. On Some Bounds for the Gamma Function. Symmetry 2023, 15, 937. https://doi.org/10.3390/sym15040937

AMA Style

Mahmoud M, Alsulami SM, Almarashi S. On Some Bounds for the Gamma Function. Symmetry. 2023; 15(4):937. https://doi.org/10.3390/sym15040937

Chicago/Turabian Style

Mahmoud, Mansour, Saud M. Alsulami, and Safiah Almarashi. 2023. "On Some Bounds for the Gamma Function" Symmetry 15, no. 4: 937. https://doi.org/10.3390/sym15040937

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop