1. Introduction

For , the power mean of order of two positive numbers and is defined by

(11)

Recently, the power mean has been the subject of intensive research. In particular, many remarkable inequalities for can be found in literature  [1–12]. It is well known that is continuous and increasing with respect to for fixed and . If we denote by and the arithmetic mean, geometric mean and harmonic mean of and , respectively, then

(12)

In [13], Alzer and Janous established the following sharp double-inequality (see also   [14,page 350]):

(13)

for all

In [15], Mao proved

(14)

for all , and is the best possible lower power mean bound for the sum .

The purpose of this paper is to answer the questions: what are the greatest values and , and the least values and , such that and for all ?

2. Main Results

Theorem 2.1.

for all , equality holds if and only if , and is the best possible lower power mean bound for the sum .

Proof.

If , then we clearly see that .

If and , then simple computation leads to

(21)

Next, we prove that is the best possible lower power mean bound for the sum .

For any and , one has

(22)

where .

Let , then the Taylor expansion leads to

(23)

Equations (2.2) and (2.3) imply that for any there exists , such that for .

Remark 2.2.

For any , one has

(24)

Therefore, is the best possible upper power mean bound for the sum .

Theorem 2.3.

for all , equality holds if and only if , and is the best possible lower power mean bound for the sum .

Proof.

If , then we clearly see that

If and , then elementary calculation yields

(25)

Next, we prove that is the best possible lower power mean bound for the sum .

For any and , one has

(26)

where .

Let , then the Taylor expansion leads to

(27)

Equations (2.6) and (2.7) imply that for any there exists , such that

(28)

for .

Remark 2.4.

For any , one has

(29)

Therefore, is the best possible upper power mean bound for the sum .