Skip to main content
Springer logoLink to Springer
. 2018 Jun 27;2018(1):147. doi: 10.1186/s13660-018-1741-8

Inequalities and asymptotic expansions related to the generalized Somos quadratic recurrence constant

Xue-Si Ma 1, Chao-Ping Chen 1,
PMCID: PMC6019429  PMID: 30008535

Abstract

In this paper, we give asymptotic expansions and inequalities related to the generalized Somos quadratic recurrence constant, using its relation with the generalized Euler constant.

Keywords: Somos’ quadratic recurrence constant, Inequality, Asymptotic expansion, Generalized Euler constant

Introduction

Somos’ quadratic recurrence constant is defined (see [13]) by

σ=123=n=1n1/2n=k=1(1+1k)1/2k=exp{k=1lnk2k}=1.66168794 1.1

or

σ=exp{011x(2x)lnxdx}=exp{0101x(2xy)ln(xy)dxdy}. 1.2

The constant σ arises in the study of the asymptotic behavior of the sequence

g0=1,gn=ngn12,nN:={1,2,3,}, 1.3

with the first few terms

g0=1,g1=1,g2=2,g3=12,g4=576,g5=1,658,880,.

This sequence behaves as follows (see [4, p. 446] and [3, 5]):

gnσ2nn(1+2n1n2+4n321n4+138n51091n6+10,088n7106,918n8+1,279,220n917,070,418n10+251,560,472n114,059,954,946n12+)1. 1.4

The constant σ appears in important problems from pure and applied analysis, and it is the motivation for a large number of research papers (see, for example, [1, 616]).

Sondow and Hadjicostas [15] introduced and studied the generalized-Euler-constant function γ(z), defined by

γ(z)=n=1zn1(1nlnn+1n), 1.5

where the series converges when |z|1. Pilehrood and Pilehrood [13] considered the function zγ(z) (|z|1). The function γ(z) generalizes both Euler’s constant γ(1) and the alternating Euler constant ln4π=γ(1) [17, 18].

Sondow and Hadjicostas [15] defined the generalized Somos constant

σt=1234tttt=n=1n1/tn=(tt1)1/(t1)exp{1t(t1)γ(1t)},t>1. 1.6

Coffey [19] gave the integral and series representations for lnσt:

lnσt=0(ext1+11tex)dxx 1.7

and

lnσt=1t1k=1(1)k1kLik(1t)=1t1k=11k[tLik(1t)1] 1.8

in terms of the polylogarithm function.

It is known (see [15]) that

γ(12)=2ln2σ,equivalently,σ=2exp{12γ(12)}. 1.9

Thus, formula (1.5) is closely related to Somos’ quadratic recurrence constant σ.

Define

γn(z)=k=1nzk1(1klnk+1k),|z|1.

Mortici [11] proved that for nN,

270(n+1)2n(270n3+1530n2+1065n+6293)<γ(12)γn(12)<182n(18n2+84n13) 1.10

and

k=n+112k23k122,400(n+1)3n(44,800n4+280,000n3+435,120n2+744,380n2,477,677)<γ(13)γn(13)<k=n+112k23k11603n(320n3+1680n2+1428n+3889). 1.11

Lu and Song [10] improved Mortici’s results and obtained the inequalities:

690n2+3524n+1456(2n)(n+1)2(115n2+894n+779)<γ(12)γn(12)<48n+1273(2n)(16n+85)(n+1)2 1.12

and

k=n+112k23k1987840n2+8444340n+1094677940(3n)(n+1)2(49392n3+582741n2+1769516n+1236167)<γ(13)γn(13)<k=n+112k23k11620n2+6995n+184740(3n)(81n2+532n+451)(n+1)3 1.13

for nN.

You and Chen [16] further improved inequalities (1.10)–(1.13). Recently, Chen and Han [7] gave new bounds for γ(1/2)γn(1/2):

12n(1(n+1)283(n+1)3+232(n+1)43325(n+1)5+479(n+1)629,0247(n+1)7)<γ(12)γn(12)<12n(1(n+1)283(n+1)3+232(n+1)43325(n+1)5+479(n+1)6) 1.14

for nN, and presented the following asymptotic expansion:

γ(12)γn(12)12n(n+1)2{183(n+1)+232(n+1)23325(n+1)3+479(n+1)4} 1.15

as n. Moreover, these authors gave a formula for successively determining the coefficients in (1.15).

Chen and Han [7] pointed out that the lower bound in (1.14) is for n24 sharper than the one in (1.12), and the upper bound in (1.14) is for n18 sharper than the one in (1.12),

For any positive integer m2, in this paper we give the asymptotic expansion of γ(1/m)γn(1/m) as n. Based on the result obtained, we establish the inequality for γ(1/4)γn(1/4). We also consider the asymptotic expansion for γ(1)γn(1).

Lemmas

Lemma 2.1

As x,

1xln(1+1x)j=2m1(1)jj1xjA(x)1mA(x+1), 2.1

where A(x) is defined by

A(x)=j=majxj 2.2

with the coefficients aj given by the recurrence relation

aj=(1)jm1{mj+k=mj1(1)kak(j1jk)},jm. 2.3

Here, and throughout this paper, an empty sum is understood to be zero.

Proof

Using the Maclaurin series of ln(1+t),

ln(1+t)=j=1(1)j1jtj,1<t1,

we obtain

1xln(1+1x)j=2m1(1)jj1xj=j=m(1)jj1xj. 2.4

In view of (2.4), we can let

1xln(1+1x)j=2m1(1)jj1xjj=majxj1mj=maj(x+1)j, 2.5

where aj are real numbers to be determined.

Write (2.5) as

1xln(1+1x)j=2m1(1)jj1xjj=majxj1mj=majxj(1+1x)j. 2.6

Direct computation yields

j=majxj(1+1x)j=j=majxjk=0(jk)1xk=j=majxjk=0(1)k(k+j1k)1xk=j=mk=mjak(1)jk(j1jk)1xj. 2.7

It follows from (2.4), (2.6), and (2.7) that

j=m(1)jj1xjj=m{aj1mk=mjak(1)jk(j1jk)}1xj. 2.8

Equating coefficients of the term xj on both sides of (2.8) yields

(1)jj=aj1mk=mjak(1)jk(j1jk),jm. 2.9

For j=m, we obtain am=(1)mm1, and for jm+1, we have

(1)jj=aj1m[k=mj1ak(1)jk(j1jk)+aj],jm+1.

We then obtain the recursive formula

am=(1)mm1,aj=(1)jm(m1)j+1m1k=mj1ak(1)jk(j1jk),jm+1,

which can be written as (2.3). The proof of Lemma 2.1 is complete. □

Lemma 2.2

Let

a(x)=13x43245x5andb(x)=13x43245x5+6827x6. 2.10

Then, for x1,

a(x)14a(x+1)<1xln(1+1x)12x2+13x3<b(x)14b(x+1). 2.11

Proof

It is well known that for 1<t1 and mN,

j=12m(1)j1jtj<ln(1+t)<j=12m1(1)j1jtj,

which implies that for x1 and m2,

j=42m+1(1)jjxj<1xln(1+1x)12x2+13x3<j=42m(1)jjxj. 2.12

Using (2.12), we find that

1xln(1+1x)12x2+13x3a(x)+14a(x+1)>14x415x5a(x)+14a(x+1)=310x4+770x3+845x2+445x+92180x5(x+1)5>0

and

1xln(1+1x)12x2+13x3b(x)+14b(x+1)<14x415x5+16x6b(x)+14b(x+1)=4380x5+14,205x4+21,530x3+17,439x2+7344x+1270540x6(x+1)6<0.

The proof of Lemma 2.2 is complete. □

Remark 2.1

Using the methods from [2022] it is possible to get estimations (based on the power series expansions) of the logarithm function that can be used, for example, in the analysis of parameterized Euler-constant function, which will be an item for further work.

Lemma 2.3

As x, we have

1xln(1+1x)C(x)+C(x+1), 2.13

where C(x) is defined by

C(x)=j=2cjxj 2.14

with the coefficients cj given by the recurrence relation

c2=14,cj=(1)j2j12k=2j1ck(1)jk(j1jk),j3. 2.15

Proof

In view of (2.4), we can let

1xln(1+1x)j=2cjxj+j=2cj(x+1)j, 2.16

where cj are real numbers to be determined. Write (2.16) as

1xln(1+1x)j=2cjxj+j=2cjxj(1+1x)j.

Noting that (2.7) holds, we have

j=2(1)jj1xjj=2{cj+k=2jck(1)jk(j1jk)}1xj. 2.17

Equating coefficients of the term xj on both sides of (2.17) yields

(1)jj=cj+k=2jck(1)jk(j1jk),j2.

For j=2, we obtain c2=1/4, and for j3 we have

(1)jj=2cj+k=2j1ck(1)jk(j1jk),j3.

We then obtain the recursive formula (2.15). The proof of Lemma 2.3 is complete. □

The first few coefficients cj are

c2=14,c3=112,c4=18,c5=110,c6=14,c7=1756,c8=1716,c9=3118.

Main results

For any positive integer m2, Theorem 3.1 gives the asymptotic expansion of γ(1/m)γn(1/m) as n.

Theorem 3.1

For any positive integer m2, we have

γ(1m)γn(1m)k=n+1j=2m1(1)jjmk11kj+A(n+1)mn,n, 3.1

where A(x) is given in (2.2). Namely,

γ(1m)γn(1m)k=n+1j=2m1(1)jjmk11kj+(1)mmn{1(m1)(n+1)m2m2(m+1)(m1)2(n+1)m+1+m2(m2+8m+3)2(m+2)(m1)3(n+1)m+2(m+1)(m3+12m2+51m+8)m26(m1)4(m+3)(n+1)m+3+m2(m6+25m5+216m4+866m3+1241m2+501m+30)24(m1)5(m+4)(n+1)m+4}. 3.2

Proof

Write (2.1) as

1kln(1+1k)j=2m1(1)jj1kj=AN(k)1mAN(k+1)+O(1kN+1), 3.3

where

AN(k)=j=mNajkj 3.4

with the coefficients aj given by the recurrence relation (2.3). From (3.3), we have

1mk1(1kln(1+1k))1mk1j=2m1(1)jj1kj=AN(k)mk1AN(k+1)mk+O(1mk1kN+1). 3.5

Adding (3.5) from k=n+1 to k=, we have

γ(1m)γn(1m)k=n+1j=2m1(1)jjmk11kj=1mn{AN(n+1)+O(1(n+1)N+1)},

which can be written as (3.1). The proof of Theorem 3.1 is complete. □

Remark 3.1

For m=2 in (3.2), we obtain (1.15). For m=3 in (3.2), we find

γ(13)γn(13)k=n+112k23k1+13n(n+1)3{12+98(n+1)8120(n+1)2+372(n+1)3566156(n+1)4+}. 3.6

For m=4 in (3.2), we find

γ(14)γn(14)k=n+1(1213k)1k24k1+14n(n+1)4{133245(n+1)+6827(n+1)22080189(n+1)3+9017162(n+1)4}. 3.7

Formula (3.7) motivated us to establish Theorem 3.2.

Theorem 3.2

For nN,

k=n+1(1213k)1k24k1+14n{13(n+1)43245(n+1)5}<γ(14)γn(14)<k=n+1(1213k)1k24k1+14n{13(n+1)43245(n+1)5+6827(n+1)6}. 3.8

Proof

From the double inequality (2.11), we have

a(k)4k1a(k+1)4k<14k1(1kln(1+1k))14k1(12k213k3)<b(k)4k1b(k+1)4k, 3.9

where a(x) and b(x) are given in (2.10). Adding inequalities (3.9) from k=n+1 to k=, we have

a(n+1)4n<k=n+114k1(1kln(1+1k))k=n+114k1(12k213k3)<b(n+1)4n,

which can be written as (3.8). The proof of Theorem 3.2 is complete. □

Remark 3.2

Inequality (3.8) can be further refined by inserting additional terms on both sides of the inequality. For example, for nN, we have

k=n+1(1213k)1k24k1+14n{13(n+1)43245(n+1)5+6827(n+1)62080189(n+1)7}<γ(14)γn(14)<k=n+1(1213k)1k24k1+14n{13(n+1)43245(n+1)5+6827(n+1)62080189(n+1)7+9017162(n+1)8}. 3.10

Remark 3.3

Following the same method as the one used in the proof of Theorem 3.2, we can prove the following inequality:

k=n+112k23k1+13n{12(n+1)3+98(n+1)48120(n+1)5+372(n+1)6566156(n+1)7}<γ(13)γn(13)<k=n+112k23k1+13n{12(n+1)3+98(n+1)48120(n+1)5+372(n+1)6} 3.11

for nN. We omit the proof.

In view of (1.14), (3.11), (3.8), and (3.10), we pose the following conjecture.

Conjecture 3.1

For any positive integer m2, we have

(1)mmnj=m2Naj(n+1)j<(1)m{γ(1m)γn(1m)k=n+1j=2m1(1)jjmk1kj}<(1)mmnj=m2N+1aj(n+1)j, 3.12

with the coefficients aj given in (2.3).

By using the Maple software, we find, as n,

γ(12)γn(12)12n(n+1)2(1+83n+8516+2689160(n+807,797129,072)3+), 3.13
γ(13)γn(13)k=n+112k23k1+13n(n+1)3(12+98n+235+9825(n+140,84327,440)3+) 3.14

and

γ(14)γn(14)k=n+1(1213k)1k24k1+14n(n+1)4{13+3245n+10924+23651134(n+825,361170,280)3+}. 3.15

From a computational viewpoint, formulas (3.13), (3.14), and (3.15) improve formulas (1.15), (3.6), and (3.7), respectively.

For any positive integer m2, we here provide a pair of recurrence relations for determining the constants pp(m) and qq(m) (see Remark 3.4) such that

γ(1m)γn(1m)k=n+1j=2m1(1)jjmk11kj+1mn(n+1)m(am+=1p(n+q)21) 3.16

as n. This develops formulas (3.13), (3.14), and (3.15) to produce a general result given by Theorem 3.3.

Theorem 3.3

For any positive integer m2, we have

γ(1m)γn1(1m)k=nj=2m1(1)jjmk11kj+1mn1nm(am+=1λ(n+μ)21) 3.17

as n, where λλ(m) and μμ(m) are given by a pair of recurrence relations

λ=am+21k=11λkμk22k(2222k),2, 3.18

and

μ=1(21)λ{am+2+k=11λkμk22k+1(2122k+1)},2, 3.19

with

λ1=am+1=(1)m+12m2(m+1)(m1)2andμ1=am+2λ1=(m+1)(m2+8m+3)4(m+2)(m1).

Here aj are given in (2.3).

Proof

In view of (3.13), (3.14), and (3.15), we let

γ(1m)γn1(1m)k=nj=2m1(1)jjmk11kj+1mn1nm(am+=1λ(n+μ)21),

where λ and μ are real numbers to be determined. This can be written as

mn1nm{γ(1m)γn1(1m)k=nj=2m1(1)jjmk11kj}am+j=1λjn2j1(1+μjn)2j+1. 3.20

Direct computation yields

j=1λjn2j1(1+μjn)2j+1=j=1λjn2j1k=0(2j+1k)μjknk=j=1λjn2j1k=0(1)k(k+2j2k)μjknk=j=1k=0j1λk+1μk+1jk1(1)jk1(j+k1jk1)1nj+k,

which can be written as

j=1λjn2j1(1+μjn)2j+1j=1{k=1j+22λkμkj2k+1(1)j1(j1j2k+1)}1nj. 3.21

Substituting (3.21) into (3.20) we have

mn1nm{γ(1m)γn1(1m)k=nj=2m1(1)jjmk11kj}am+j=1{k=1j+22λkμkj2k+1(1)j1(j1j2k+1)}1nj. 3.22

On the other hand, it follows from (3.1) that

mn1nm{γ(1m)γn1(1m)k=nj=2m1(1)jjmk11kj}j=0am+jnj. 3.23

Equating coefficients of the term nj on the right-hand sides of (3.22) and (3.23), we obtain

am+j=k=1j+22λkμkj2k+1(1)j1(j1j2k+1),jN. 3.24

Setting j=21 and j=2 in (3.24), respectively, yields

am+21=k=1λkμk22k(2222k) 3.25

and

am+2=k=1+1λkμk22k+1(2122k+1)=k=1λkμk22k+1(2122k+1)λ+1μ+11(211)=k=1λkμk22k+1(2122k+1). 3.26

For =1, from (3.25) and (3.26) we obtain

λ1=am+1=(1)m+12m2(m+1)(m1)2andμ1=am+2λ1=(m+1)(m2+8m+3)4(m+2)(m1),

and for 2 we have

am+21=k=11λkμk22k(2222k)+λ

and

am+2=k=11λkμk22k+1(2122k+1)(21)λμ.

We then obtain the recurrence relations (3.18) and (3.19). The proof of Theorem 3.3 is complete. □

Here we give explicit numerical values of some first terms of λ and μ by using formulas (3.18) and (3.19). This shows how easily we can determine the constants a and b in (3.17).

λ1=am+1=(1)m+12m2(m+1)(m1)2,μ1=am+2λ1=(m+1)(m2+8m+3)4(m+2)(m1),λ2=am+3λ1μ12=(1)m+1m2(m+1)(m3+12m2+51m+8)6(m1)4(m+3)(1)m+12m2(m+1)(m1)2((m+1)(m2+8m+3)4(m+2)(m1))2=(1)m+1m2(m+1)(m5+7m4+58m3+266m2+485m+47)24(m1)4(m+3)(m+2)2,μ2=am+4+λ1μ133λ2=(1)mm2(m6+25m5+216m4+866m3+1241m2+501m+30)24(m1)5(m+4)+λ1μ133λ2=((m+3)(m9+34m8+450m7+3634m6+17,584m5+48,642m4+71,302m3+50,926m2+14,151m+636))/(12(m+2)(m+4)(m5+7m4+58m3+266m2+485m+47)(m21)).

Remark 3.4

The constants p and q in (3.16) are given by

p:=λandq:=1+μ.

Setting m=2,3, and 4 in (3.16), respectively, yields (3.13), (3.14), and (3.15).

Noting that ln4π=γ(1) holds, Theorem 3.4 presents the asymptotic expansion for ln4π.

Theorem 3.4

As n, we have

γ(1)γn(1)(1)nC(n+1), 3.27

where C(x) is given in (2.14). Namely,

γ(1)γn(1)(1)n{14(n+1)2+112(n+1)318(n+1)4110(n+1)5+}. 3.28

Proof

Write (2.13) as

1kln(1+1k)=CN(k)+CN(k+1)+O(1kN+1), 3.29

where

CN(x)=j=2Ncjxj 3.30

with the coefficients cj given by the recurrence relation (2.15).

From (3.29), we have

(1)k1(1kln(1+1k))=(1)k1CN(k)(1)kCN(k+1)+O((1)k1kN+1). 3.31

Adding (3.31) from k=n+1 to k=, we have

γ(1)γn(1)=k=n+1(1)k1(1kln(1+1k))=(1)nCN(n+1)+O(1(n+1)N+1), 3.32

which can be written as (3.27). The proof of Theorem 3.4 is complete. □

Remark 3.5

We see from (3.28) that the alternating Euler constant ln4π has the following expansion:

ln4πk=1n(1)k1(1klnk+1k)+(1)n{14(n+1)2+112(n+1)318(n+1)4110(n+1)5+}. 3.33

Conclusions

In this paper, we give asymptotic expansions related to the generalized Somos quadratic recurrence constant (Theorems 3.1 and 3.3). We present the inequalities for γ(14)γn(14) and γ(13)γn(13) (see (3.8), (3.10), and (3.11)). The expansion of the alternating Euler constant ln4π is also obtained (see (3.33)).

Acknowledgements

We thank the editor and referees for their careful reading and valuable suggestions to make the article reader friendly.

Authors’ contributions

Both authors contributed equally to this work. Both authors read and approved the final manuscript.

Funding

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Footnotes

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Contributor Information

Xue-Si Ma, Email: hpumxs@sina.com.

Chao-Ping Chen, Email: chenchaoping@sohu.com.

References

  • 1.Guillera J., Sondow J. Double integrals and infinite products for some classical constants via analytic continuations of Lerch’s transcendent. Ramanujan J. 2008;16:247–270. doi: 10.1007/s11139-007-9102-0. [DOI] [Google Scholar]
  • 2.Ramanujan S. In: Collected Papers of Srinivasa Ramanujan. Hardy G.H., Aiyar P.V.S., Wilson B.M., editors. Providence: Am. Math. Soc.; 2000. [Google Scholar]
  • 3. Sloane, N.J.A.: Sequences A116603. In: The On-Line Encyclopedia of Integer Sequences. http://oeis.org/A116603
  • 4.Finch S.R. Mathematical Constants. Cambridge: Cambridge University Press; 2003. [Google Scholar]
  • 5. Weisstein, E.W.: Somos’s quadratic recurrence constant. In: MathWorld—A Wolfram Web Resource. Published electronically at http://mathworld.wolfram.com/SomossQuadraticRecurrenceConstant.html
  • 6.Chen C.P. New asymptotic expansions related to Somos’ quadratic recurrence constant. C. R. Acad. Sci. Paris, Ser. I. 2013;351:9–12. doi: 10.1016/j.crma.2013.01.005. [DOI] [Google Scholar]
  • 7.Chen C.P., Han X.F. On Somos’ quadratic recurrence constant. J. Number Theory. 2016;166:31–40. doi: 10.1016/j.jnt.2016.02.018. [DOI] [Google Scholar]
  • 8.Hirschhorn M.D. A note on Somos’ quadratic recurrence constant. J. Number Theory. 2011;131:2061–2063. doi: 10.1016/j.jnt.2011.04.010. [DOI] [Google Scholar]
  • 9.Lampret V. Approximation of Sondow’s generalized-Euler-constant function on the interval [1,1] Ann. Univ. Ferrara. 2010;56:65–76. doi: 10.1007/s11565-009-0089-x. [DOI] [Google Scholar]
  • 10.Lu D., Song Z. Some new continued fraction estimates of the Somos’ quadratic recurrence constant. J. Number Theory. 2015;155:36–45. doi: 10.1016/j.jnt.2015.03.013. [DOI] [Google Scholar]
  • 11.Mortici C. Estimating the Somos’ quadratic recurrence constant. J. Number Theory. 2010;130:2650–2657. doi: 10.1016/j.jnt.2010.06.012. [DOI] [Google Scholar]
  • 12.Nemes G. On the coefficients of an asymptotic expansion related to Somos’ quadratic recurrence constant. Appl. Anal. Discrete Math. 2011;5:60–66. doi: 10.2298/AADM101115029N. [DOI] [Google Scholar]
  • 13.Pilehrood K.H., Pilehrood T.H. Arithmetical properties of some series with logarithmic coefficients. Math. Z. 2007;255:117–131. doi: 10.1007/s00209-006-0015-1. [DOI] [Google Scholar]
  • 14.Pilehrood K.H., Pilehrood T.H. Vacca-type series for values of the generalized Euler constant function and its derivative. J. Integer Seq. 2010;13:10.7.3. [Google Scholar]
  • 15.Sondow J., Hadjicostas P. The generalized-Euler-constant function γ(z) and a generalization of Somos’s quadratic recurrence constant. J. Math. Anal. Appl. 2007;332:292–314. doi: 10.1016/j.jmaa.2006.09.081. [DOI] [Google Scholar]
  • 16.You X., Chen D.R. Improved continued fraction sequence convergent to the Somos’ quadratic recurrence constant. J. Math. Anal. Appl. 2016;436:513–520. doi: 10.1016/j.jmaa.2015.12.013. [DOI] [Google Scholar]
  • 17.Sondow J. Double integrals for Euler’s constant and ln(4/π) and an analog of Hadjicostas’s formula. Am. Math. Mon. 2005;112:61–65. [Google Scholar]
  • 18.Sondow J. Additive Number Theory. New York: Springer; 2010. New Vacca-type rational series for Euler’s constant and its “alternating” analog ln(4/π) pp. 331–340. [Google Scholar]
  • 19.Coffey M.W. Integral representations of functions and Addison-type series for mathematical constants. J. Number Theory. 2015;157:79–98. doi: 10.1016/j.jnt.2015.04.005. [DOI] [Google Scholar]
  • 20. Lutovac, T., Malesevic, B., Rasajski, M.: A new method for proving some inequalities related to several special functions (2018). arXiv:1802.02082
  • 21.Malesevic B., Lutovac T., Rasajski M., Mortici C. Extensions of the natural approach to refinements and generalizations of some trigonometric inequalities. Adv. Differ. Equ. 2018;2018:90. doi: 10.1186/s13662-018-1545-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Malesevic B., Rasajski M., Lutovac T. Refinements and generalizations of some inequalities of Shafer-Fink’s type for the inverse sine function. J. Inequal. Appl. 2017;2017:275. doi: 10.1186/s13660-017-1554-1. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Journal of Inequalities and Applications are provided here courtesy of Springer

RESOURCES