There are many relations involving the geometric means $G_n\left(x\right)$ and power means ${\left[A_n\left(x^{\gamma }\right)\right]}^{1/\gamma }$ for positive $n$-vectors $x$. Some of them assume the form of inequalities involving parameters. There then is the question of sharpness, which is quite difficult in general. In this paper we are concerned with inequalities of the form $(1-\lambda )G_n^{\gamma }\left(x\right)+\lambda A_n^{\gamma }\left(x\right)\ge A_n\left(x^{\gamma }\right)$ and $(1-\lambda )G_n^{\gamma }\left(x\right)+\lambda A_n^{\gamma }\left(x\right)\le A_n\left(x^{\gamma }\right)$ with parameters $\lambda \in ℝ$ and $\gamma \in (0,1).$ We obtain a necessary and sufficient condition for the former inequality, and a sharp condition for the latter. Several applications of our results are also demonstrated....

We provide a set of optimal estimates of the form (1-μ)/𝓐(x,y) + μ/ℳ (x,y) ≤ 1/ℬ(x,y) ≤ (1-ν)/𝓐(x,y) + ν/ℳ (x,y) where 𝓐 < ℬ are two of the Seiffert means L,P,M,T, while ℳ is another mean greater than the two.