1 Notation and preliminaries

The setting for this paper is an n-dimensional Euclidean space \({\Bbb {R}}^{n}\). We reserve the letter u for unit vectors, and the letter B is reserved for the unit ball centered at the origin. The surface of B is \(S^{n-1}\). The volume of the unit n-ball is denoted by \(\omega_{n}\). We use \(V(K)\) for the n-dimensional volume of a body K. Associated with a compact subset K of \({\Bbb {R}}^{n}\), which is star-shaped with respect to the origin, is its radial function \(\rho(K,\cdot): S^{n-1}\rightarrow{\Bbb {R}}\) defined for \(u\in S^{n-1}\) by

$$\rho(K,u)=\max\{\lambda\geq0: \lambda u\in K\}. $$

If \(\rho(K,\cdot)\) is positive and continuous, K will be called a star body. Let \({\mathcal {S}}^{n}\) denote the set of star bodies in \({\Bbb {R}}^{n}\). Let δ̃ denote the radial Hausdorff metric, i.e., if \(K, L\in{\mathcal {S}}^{n}\), then \(\tilde{\delta}(K,L)=|\rho(K,u)-\rho(L,u)|_{\infty}\), where \(|\cdot|_{\infty}\) denotes the sup-norm on the space of continuous functions \(C(S^{n-1})\).

1.1 Dual mixed volumes

The radial Minkowski linear combination, \(\lambda_{1}K_{1}\widetilde {+}\cdots\widetilde{+}\lambda_{r} K_{r}\) is defined by

$$\begin{aligned} \lambda_{1}K_{1}\widetilde{+}\cdots\widetilde{+} \lambda_{r} K_{r}=\{\lambda_{1}x_{1} \widetilde{+}\cdots\widetilde{+}\lambda_{r} x_{r}: x_{i}\in K_{i}, i=1,\ldots,r\} \end{aligned}$$
(1.1)

for \(K_{1},\ldots,K_{r}\in{\mathcal {S}}^{n}\) and \(\lambda_{1},\ldots,\lambda_{r}\in{\Bbb {R}}\). It has the following important property (see [1]):

$$\begin{aligned} \rho(\lambda K\widetilde{+}\mu L,\cdot)=\lambda\rho(K,\cdot)+\mu\rho(L, \cdot) \end{aligned}$$
(1.2)

for \(K, L\in{\mathcal {S}}^{n}\) and \(\lambda, \mu\geq0\). For \(K_{1},\ldots,K_{r}\in{\mathcal {S}}^{n}\) and \(\lambda_{1},\ldots,\lambda_{r}\geq0\), the volume of the radial Minkowski linear combination \(\lambda_{1}K_{1}\widetilde{+}\cdots\widetilde{+}\lambda_{r}K_{r}\) is a homogeneous polynomial of degree n in the \(\lambda_{i}\),

$$\begin{aligned} V(\lambda_{1}K_{1}\widetilde{+}\cdots\widetilde{+}\lambda _{r}K_{r})=\sum_{i_{1},\ldots,i_{n}=1}^{r} \widetilde{V}(K_{i_{1}},\ldots,K_{i_{n}})\lambda_{i_{1}} \cdots \lambda_{i_{n}}. \end{aligned}$$
(1.3)

If we require the coefficients of the polynomial in (1.3) to be symmetric in their arguments, then they are uniquely determined. The coefficient \(\widetilde{V}(K_{i_{1}},\ldots,K_{i_{n}})\) is nonnegative and depends only on the bodies \(K_{i_{1}},\ldots,K_{i_{n}}\). It is called the dual mixed volume of \(K_{i_{1}},\ldots,K_{i_{n}}\).

If \(K_{1},\ldots,K_{n}\in{\mathcal {S}}^{n}\), then the dual mixed volume \(\widetilde{V}(K_{1},\ldots,K_{n})\) can be represented in the form (see [2])

$$\begin{aligned} \widetilde{V}(K_{1},\ldots,K_{n})=\frac{1}{n} \int_{S^{n-1}}\rho (K_{1},u)\cdots\rho(K_{n},u)\,dS(u). \end{aligned}$$
(1.4)

If \(K_{1}=\cdots=K_{n-i}=K\), \(K_{n-i+1}=\cdots=K_{n}=L\), then the dual mixed volume is written as \(\widetilde{V}_{i}(K,L)\). If \(L=B\), then the dual mixed volume \(\widetilde{V}_{i}(K,L)=\widetilde{V}_{i}(K,B)\) is written as \(\widetilde{W}_{i}(K)\). For \(K,L\in{\mathcal {S}}^{n}\), the ith dual mixed volume of K and L, \(\widetilde{V}_{i}(K,L)\) can be extended to all \(i\in{\Bbb {R}}\) by

$$\begin{aligned} \widetilde{V}_{i}(K,L)=\frac{1}{n} \int_{S^{n-1}}\rho(K,u)^{n-i}\rho (L,u)^{i}\,dS(u), \end{aligned}$$
(1.5)

where \(i\in{\Bbb {R}}\). Thus, if \(K\in{\mathcal {S}}^{n}\), then

$$\begin{aligned} \widetilde{W}_{i}(K)=\frac{1}{n} \int_{S^{n-1}}\rho (K,u)^{n-i}\,dS(u). \end{aligned}$$
(1.6)

1.2 Mixed intersection bodies

For \(K\in{\mathcal {S}}^{n}\), there is a unique star body I K whose radial function satisfies, for \(u\in S^{n-1}\),

$$\rho(\mathbf{I}K,u)=v(K\cap E_{u}), $$

where v is \((n-1)\)-dimensional dual volume. It is called the intersection body of K. The volume of the intersection body of K is given by (see [1])

$$V(\mathbf{I}K)=\frac{1}{n} \int_{S^{n-1}}v(K\cap E_{u})^{n}\,dS(u). $$

The mixed intersection body of \(K_{1},\ldots,K_{n-1}\in{\mathcal {S} }^{n}\), denoted by \(\mathbf{I}(K_{1},\ldots,K_{n-1})\), is defined by

$$\rho\bigl(\mathbf{I}(K_{1},\ldots,K_{n-1}),u\bigr)= \tilde{v}(K_{1}\cap E_{u},\ldots,K_{n-1}\cap E_{u}), $$

where is the \((n-1)\)-dimensional dual mixed volume. If \(K_{1}=\cdots=K_{n-i-1}=K, K_{n-i}=\cdots=K_{n-1}=L\), then \({\bf I}(K_{1},\ldots,K_{n-1})\) is written as \(\mathbf{I}_{i}(K,L)\). If \(L=B\), then \(\mathbf{I}_{i}(K,L)\) is written as \(\mathbf{I}_{i}K\) and called the ith intersection body of K. For \(\mathbf{I}_{0}K\), we simply write I K.

2 Improvement of the radial Blaschke addition

Let us recall the concept of radial Blaschke addition defined by Lutwak [1]. Suppose that K and L are star bodies in \({\Bbb {R}}^{n}\), the radial Blaschke addition denoted by \(K\widehat{+}L\) is a star body whose radial function is

$$\begin{aligned} \rho(K\widehat{+}L, \cdot)^{n-1}=\rho(K,\cdot)^{n-1}+\rho(L, \cdot)^{n-1}. \end{aligned}$$
(2.1)

The dual Knesser-Süss inequality for the radial Blaschke addition was established by Lutwak [1]. If \(K,L\in{\mathcal {S}}^{n}\), then

$$\begin{aligned} V(K\widehat{+}L)^{(n-1)/n}\leq V(K)^{(n-1)/n}+V(L)^{(n-1)/n}, \end{aligned}$$
(2.2)

with equality if and only if K and L are dilates.

In the section, we give a generalized concept of the radial Blaschke addition.

Definition 2.1

If \(K,L\in{\mathcal {S}}^{n}\), \(0\leq p< n-1\) and \(\lambda, \mu>0\) (not both zero), the p-radial Blaschke linear combination of K and L denoted by \(\lambda\diamond K\widehat{+}_{p}\mu\diamond L\) is a star body whose radial function is defined by

$$\begin{aligned} \rho(\lambda\diamond K\widehat{+}_{p}\mu\diamond L,\cdot )^{n-p-1}=\lambda\rho(K,\cdot)^{n-p-1}+\mu\rho(L,\cdot )^{n-p-1}. \end{aligned}$$
(2.3)

From (2.3), it is easy to see that

$$\lambda\diamond K=\lambda^{1/(n-p-1)} K. $$

When \(\lambda=\mu=1\), the p-radial Blaschke combination becomes the p-radial Blaschke addition \(K\widehat{+}_{p}L\) and

$$\begin{aligned} \rho(K\widehat{+}_{p}L,\cdot)^{n-p-1}=\rho(K, \cdot)^{n-p-1}+\rho (L,\cdot)^{n-p-1}. \end{aligned}$$
(2.4)

Obviously, when \(p=0\), (2.4) becomes (2.1).

In the following, we define the dual mixed quermassintegral with respect to the p-radial Blaschke addition. First, we show two propositions. The following proposition follows immediately from (2.3) with L’Hôpital’s rule.

Proposition 2.2

Let \(0\leq p< n-1\), \(0\leq i< n\) and \(\varepsilon>0\). If \(K,L\in{\mathcal {S}}^{n}\), then

$$\begin{aligned} \lim_{\varepsilon\rightarrow0^{+}}\frac{\rho(K\widehat {+}_{p}\varepsilon\diamond L,u)^{n-i}-\rho(K,u)^{n-i}}{\varepsilon}=\frac{n-i}{n-p-1}\rho (K,u)^{p-i+1}\rho(L,u)^{n-p-1}. \end{aligned}$$
(2.5)

The following proposition follows immediately from Proposition 2.2 and (1.6).

Proposition 2.3

Let \(0\leq p< n-1\), \(0\leq i< n\) and \(\varepsilon>0\). If \(K,L\in{\mathcal {S}}^{n}\), then

$$\begin{aligned} &\frac{n-p-1}{n-i}\lim_{\varepsilon\rightarrow0^{+}}\frac {\widetilde{W}_{i}(K\widehat{+}_{p}\varepsilon\diamond L,u)-\widetilde{W}_{i}(K)}{\varepsilon} \\ &\quad=\frac{1}{n} \int _{S^{n-1}}\rho(K,u)^{p-i+1}\rho(L,u)^{n-p-1}\,dS(u). \end{aligned}$$
(2.6)

Definition 2.4

For \(0\leq p< n-1\), \(0\leq i< n\) and \(K,L\in {\mathcal {S}}^{n}\), the p-dual mixed quermassintegral of star bodies K and L, denoted by \(\widetilde{W}_{p,i}(K,L)\), is defined by

$$\begin{aligned} \widetilde{W}_{p,i}(K,L)=\frac{1}{n} \int_{S^{n-1}}\rho (K,u)^{p-i+1}\rho(L,u)^{n-p-1}\,dS(u). \end{aligned}$$
(2.7)

Obviously, when \(K=L\), \(\widetilde{W}_{p,i}(K,L)\) becomes the dual quermassintegral of star body K, i.e., \(\widetilde{W}_{p,i}(K,K)=\widetilde{W}_{i}(K)\). Taking \(i=0\) in (2.7), \(\widetilde{W}_{p,i}(K,L)\) becomes the p-dual mixed volume \(\widetilde{V}_{p}(K,L)\) and

$$\begin{aligned} \widetilde{V}_{p}(K,L)=\frac{1}{n} \int_{S^{n-1}}\rho(K,u)^{p+1}\rho (L,u)^{n-p-1}\,dS(u). \end{aligned}$$
(2.8)

From (2.7), combining Hölder’s integral inequality (see [3]) gives the following.

Proposition 2.5

(Minkowski type inequality)

If \(K,L\in{\mathcal {S}}^{n}\), \(0\leq i< n\) and \(0\leq p< n-1\), then

$$\begin{aligned} \widetilde{W}_{p,i}(K,L)^{n-i}\leq\widetilde {W}_{i}(K)^{p-i+1}\widetilde{W}_{i}(L)^{n-p-1}, \end{aligned}$$
(2.9)

with equality if and only if K and L are dilates.

Taking \(i=0\) in (2.9), we have: If \(K,L\in{\mathcal {S}}^{n}\) and \(0\leq p< n-1\), then

$$\begin{aligned} \widetilde{V}_{p}(K,L)^{n}\leq V(K)^{p+1}V(L)^{n-p-1}, \end{aligned}$$
(2.10)

with equality if and only if K and L are dilates. In the following, we establish the Brunn-Minkowski inequality for the p-radial Blaschke addition.

Proposition 2.6

If \(K,L\in{\mathcal {S}}^{n}\), \(0\leq i< n\) and \(0\leq p< n-1\), then

$$\begin{aligned} \widetilde{W}_{i}(K\widehat{+}_{p}L)^{(n-p-1)/(n-i)}\leq \widetilde {W}_{i}(K)^{(n-p-1)/(n-i)}+\widetilde{W}_{i}(L)^{(n-p-1)/(n-i)}, \end{aligned}$$
(2.11)

with equality if and only if K and L are dilates.

Proof

From (2.3) and (2.7), it is easily seen that the p-dual mixed quermassintegral \(\widehat{W}_{p,i}(K,L)\) is linear with respect to the p-radial Blaschke addition and together with inequality (2.9) shows that

$$\begin{aligned} \widetilde{W}_{p,i}(Q, K\widehat{+}_{p}L)&=\widetilde {W}_{p,i}(Q,K)+\widetilde {W}_{p,i}(Q,L) \\ &\leq\widetilde{W}_{i}(Q)^{(p-i+1)/(n-i)}\bigl(\widetilde {W}_{i}(K)^{(n-p-1)/(n-i)}+\widetilde {W}_{i}(L)^{(n-p-1)/(n-i)} \bigr), \end{aligned}$$
(2.12)

with equality if and only if K and L are dilates of Q. Take \(K\widehat{+}_{p}L\) for Q in (2.12), recall that \(\widetilde{W}_{p,i}(Q,Q)=\widetilde{W}_{i}(Q)\), inequality (2.11) follows easy.

Taking \(i=0\) in (2.11), we obtain that if \(K,L\in{\mathcal {S}}^{n}\) and \(0\leq p< n-1\), then

$$V(K\widehat{+}_{p} L)^{(n-p-1)/n}\leq V(K)^{(n-p-1)/n}+V(L)^{(n-p-1)/n}, $$

with equality if and only if K and L are dilates. Taking \(p=0\) and \(i=0\) in (2.11), (2.11) becomes the well-known dual Knesser-Süss inequality (2.2). □

3 Improvement of the harmonic Blaschke addition

Let us recall the concept of harmonic Blaschke addition defined by Lutwak [4]. Suppose that K and L are star bodies in \({\Bbb {R}}^{n}\), the harmonic Blaschke addition denoted by \(K\breve{+}L\) is defined by

$$\begin{aligned} \frac{\rho(K\breve{+}L, \cdot)^{n+1}}{V(K\breve{+}L)}=\frac{\rho(K,\cdot )^{n+1}}{V(K)}+\frac{\rho(L,\cdot)^{n+1}}{V(L)}. \end{aligned}$$
(3.1)

Lutwak’s Brunn-Minkowski inequality for the harmonic Blaschke addition was established (see [4]). If \(K,L\in{\mathcal {S}}^{n}\), then

$$\begin{aligned} V(K\breve{+}L)^{1/n}\geq V(K)^{1/n}+V(L)^{1/n}, \end{aligned}$$
(3.2)

with equality if and only if K and L are dilates.

In the section, we give an improved concept of the harmonic Blaschke addition.

Definition 3.1

For \(0\leq i< n\), \(p< i-1\) and \(K,L\in{\mathcal {S}}^{n}\), we define the p-harmonic Blaschke addition of star bodies K and L denoted by \(K\breve{+}_{p}L\) and defined by

$$\begin{aligned} \frac{\rho(K\breve{+}_{p}L,\cdot)^{n-p-1}}{\tilde{W}_{i}(K\breve {+}_{p}L)}=\frac{\rho(K,\cdot)^{n-p-1}}{ \widetilde{W}_{i}(K)}+\frac{\rho(L,\cdot)^{n-p-1}}{\tilde {W}_{i}(L)}. \end{aligned}$$
(3.3)

The Brunn-Minkowski inequality for the p-harmonic Blaschke addition follows immediately from (1.6), (3.3) and Minkowski’s integral inequality (see [3]).

Proposition 3.2

If \(K,L\in{\mathcal {S}}^{n}\), \(0\leq i< n\) and \(p< i-1\), then

$$\begin{aligned} \widetilde{W}_{i}(K\breve{+}_{p}L)^{-(p+1-i)/(n-i)}\leq \widetilde {W}_{i}(K)^{-(p+1-i)/(n-i)} +\widetilde{W}_{i}(L)^{-(p+1-i)/(n-i)}, \end{aligned}$$
(3.4)

with equality if and only if K and L are dilates.

4 Radial Blaschke-Minkowski homomorphisms

Definition 4.1

([5])

A map \(\Psi: {\mathcal {S}}^{n}\rightarrow {\mathcal {S}}^{n}\) is called a radial Blaschke-Minkowski homomorphism if it satisfies the following conditions:

  1. (a)

    Ψ is continuous.

  2. (b)

    For all \(K,L\in{\mathcal {S}}^{n}\),

    $$\Psi(K\ddot{+}L)=\Psi(K)\widetilde{+}\Psi(L). $$
  3. (c)

    For all \(K,L\in{\mathcal {S}}^{n}\) and every \(\vartheta\in SO(n)\),

    $$\Psi(\vartheta K)=\vartheta\Psi(K), $$

    where \(SO(n)\) is the group of rotations in n dimensions.

Radial Blaschke-Minkowski homomorphisms are important examples of star body valued valuations. Their natural duals, Blaschke-Minkowski homomorphisms, are an important notion in the theory of convex body valued valuations (see, e.g., [612] and [1320]). In 2006, Schuster [5] established the following Brunn-Minkowski inequality for radial Blaschke-Minkowski homomorphisms of star bodies. If K and L are star bodies in \({\Bbb {R}}^{n}\), then

$$\begin{aligned} V\bigl(\Psi(K\widetilde{+}L)\bigr)^{1/n(n-1)}\leq V(\Psi K)^{1/n(n-1)}+V(\Psi L)^{1/n(n-1)}, \end{aligned}$$
(4.1)

with equality if and only if K and L are dilates.

If K and L are star bodies in \({\Bbb {R}}^{n}\), \(p\neq0\) and \(\lambda, \mu\geq0\), then \(\lambda\cdot K\widetilde{+}_{p}\mu\cdot L\) is the star body whose radial function is given by (see, e.g., [21])

$$\begin{aligned} \rho(\lambda\cdot K\widetilde{+}_{p}\mu\cdot L,\cdot)^{p}= \lambda\rho(K,\cdot)^{p}+\mu\rho(L,\cdot )^{p}. \end{aligned}$$
(4.2)

The addition \(\widetilde{+}_{p}\) is called \(L_{p}\)-radial addition. The \(L_{p}\) dual Brunn-Minkowski inequality states: If \(K,L\in{\mathcal {S}}^{n}\) and \(0< p\leq n\), then

$$V(K\widetilde{+}_{p}L)^{p/n}\leq V(K)^{p/n}+V(L)^{p/n}, $$

with equality when \(p\neq n\) if and only if K and L are dilates. The inequality is reversed when \(p>n\) or \(p<0\) (see [21]).

In 2013, an \(L_{p}\) Brunn-Minkowski inequality for radial Blaschke-Minkowski homomorphisms was established in [22]: If K and L are star bodies in \({\Bbb {R}}^{n}\) and \(0< p< n-1\), then

$$\begin{aligned} V\bigl(\Psi(K\widetilde{+}_{p}L)\bigr)^{p/n(n-1)}\leq V(\Psi K)^{p/n(n-1)}+V(\Psi L)^{p/n(n-1)}, \end{aligned}$$
(4.3)

with equality if and only if K and L are dilates. Taking \(p=1\), (4.3) reduces to (4.1).

Theorem 4.2

(see [5])

Let \(\Psi: {\mathcal {S}}^{n}\rightarrow{\mathcal {S}}^{n}\) be a radial Blaschke-Minkowski homomorphism. There is a continuous operator \(\Psi: \underbrace{{{\mathcal {S}}^{n}\times\cdots\times\mathcal {S}}^{n}}_{n-1}\rightarrow{\mathcal {S}}^{n}\) symmetric in its arguments such that, for \(K_{1},\ldots,K_{m}\in{\mathcal {S}}^{n}\) and \(\lambda_{1},\ldots,\lambda_{m}\geq0\),

$$\begin{aligned} \Psi(\lambda_{1}K_{1}\widetilde{+}\cdots\widetilde{+} \lambda _{m}K_{m})=\sum_{i_{1},\ldots,i_{n-1}} \lambda_{i_{1}}\cdots\lambda _{i_{n-1}} \Psi(K_{i_{1}}, \ldots,K_{i_{n-1}}). \end{aligned}$$
(4.4)

Clearly, Theorem 4.2 generalizes the notion of radial Blaschke-Minkowski homomorphisms. We call \(\Psi: {\mathcal {S}}^{n}\times\cdots\times{\mathcal {S}}^{n}\rightarrow{\mathcal {S}}^{n}\) a mixed radial Blaschke-Minkowski homomorphism induced by Ψ. Mixed radial Blaschke-Minkowski homomorphisms were first studied in more detail in [23, 24]. If \(K_{1}=\cdots=K_{n-i-1}=K, K_{n-i}=\cdots=K_{n-1}=L\), we write \(\Psi_{i}(K,L)\) for \(\Psi(\underbrace{K,\ldots,K}_{n-i-1},\underbrace{L,\ldots,L}_{i})\). If \(K_{1}=\cdots=K_{n-i-1}=K, K_{n-i}=\cdots=K_{n-1}=B\), we write \(\Psi_{i} K\) for \(\Psi(\underbrace{K,\ldots,K}_{n-i-1},\underbrace{B,\ldots,B}_{i})\) and call \(\Psi_{i} K\) the mixed Blaschke-Minkowski homomorphism of order i of K. \(\Psi_{0} K\) is written simply as ΨK.

Lemma 4.3

(see [5])

A map \(\Psi: {\mathcal {S}}^{n}\rightarrow{\mathcal {S}}^{n}\) is a radial Blaschke-Minkowski homomorphism if and only if there is a measure \(\mu\in{\mathcal {M}}_{+}(S^{n-1},\hat{e})\) such that

$$\begin{aligned} \rho(\Psi K,\cdot)=\rho(K,\cdot)^{n-1}\ast\mu, \end{aligned}$$
(4.5)

where \({\mathcal {M}}_{+}(S^{n-1},\hat{e})\) denotes the set of nonnegative zonal measures on \(S^{n-1}\).

For the mixed radial Blaschke-Minkowski homomorphism induced by Ψ, Schuster [5] proved that

$$\rho\bigl(\Psi(K_{1},\ldots,K_{n-1}),\cdot\bigr)= \rho(K_{1},\cdot)\cdots \rho(K_{n-1},\cdot)\ast\mu. $$

Obviously, a special case is the following:

$$\rho(\Psi_{i}K,\cdot)=\rho(K,\cdot)^{n-1-i}\ast\mu, $$

where i are integers. We now extend the integers i to real numbers, define the Blaschke-Minkowski homomorphism of order p of K.

Definition 4.4

Let \(K\in{\mathcal {S}}^{n}\), the Blaschke-Minkowski homomorphism of order p of K, denoted by \(\Psi_{p}K\), is defined for all \(p\in{\Bbb {R}}\) by

$$\begin{aligned} \rho(\Psi_{p}K,\cdot)=\rho(K,\cdot)^{n-1-p}\ast \mu. \end{aligned}$$
(4.6)

This extended definition will be required to prove our main results.

5 Inequalities for the radial Blaschke-Minkowski homomorphism

Theorem 5.1

Let \(K,L\in{\mathcal {S}}^{n}\). If \(0\leq p< n-1\) and \(i\leq n-1\leq j\leq n\), then

$$\begin{aligned} \biggl(\frac{\widetilde{W}_{i}(\Psi_{p}(K\widehat {+}_{p}L))}{\widetilde{W}_{j}(\Psi_{p}(K\widehat{+}_{p}L))} \biggr)^{1/(j-i)} \leq \biggl(\frac{\widetilde{W}_{i}(\Psi_{p}K)}{\widetilde {W}_{j}(\Psi_{p}K)} \biggr) ^{1/(j-i)}+ \biggl(\frac{\widetilde{W}_{i}(\Psi_{p} L)}{\widetilde{W}_{j}(\Psi_{p}L)} \biggr)^{1/(j-i)}, \end{aligned}$$
(5.1)

with equality if and only if \(\Psi_{p} K\) and \(\Psi_{p} L\) are dilates.

Remark 5.2

Taking \(j=n\) in (5.1) and noting that \(\widetilde{W}_{n}(K)=\int_{S^{n-1}}\,dS(u)=n\omega_{n}\), (5.1) becomes the following inequality: If \(K,L\in{\mathcal {S}}^{n}\), \(0\leq p< n-1\) and \(i\leq n-1\), then

$$\begin{aligned} \widetilde{W}_{i}\bigl(\Psi_{p}(K\widehat{+}_{p}L) \bigr)^{1/(n-i)} \leq\widetilde{W}_{i}(\Psi_{p} K)^{1/(n-i)}+\widetilde{W}_{i}(\Psi_{p}L)^{1/(n-i)}, \end{aligned}$$
(5.2)

with equality if and only if \(\Psi_{p} K\) and \(\Psi_{p} L\) are dilates. Taking \(p=0\) in (5.1), (5.1) becomes the following inequality: If \(K,L\in{\mathcal {S}}^{n}\) and \(i\leq n-1\leq j\leq n\), then

$$\begin{aligned} \biggl(\frac{\widetilde{W}_{i}(\Psi(K\widehat{+}L))}{\widetilde {W}_{j}(\Psi(K\widehat{+}L))} \biggr)^{1/(j-i)} \leq \biggl(\frac{\widetilde{W}_{i}(\Psi K)}{\widetilde{W}_{j}(\Psi K)} \biggr)^{1/(j-i)}+ \biggl(\frac{\widetilde{W}_{i}(\Psi L)}{\widetilde{W}_{j}(\Psi L)} \biggr)^{1/(j-i)}, \end{aligned}$$
(5.3)

with equality if and only if ΨK and ΨL are dilates.

Theorem 5.3

Let \(K,L\in{\mathcal {S}}^{n}\). If \(0\leq i< n \), \(p< i-1\) and \(k,j\in{\Bbb {R}}\) satisfy \(j\leq n-1\leq k\leq n\), then

$$\begin{aligned} &\frac{1}{\widetilde{W}_{i}(K\breve{+}_{p} L)} \biggl(\frac{\widetilde{W}_{j}(\Psi_{p}(K\breve {+}_{p}L))}{\widetilde{W}_{k}(\Psi_{p}(K\breve{+}_{p}L))} \biggr)^{1/(k-j)} \\ &\quad\leq\frac{1}{\widetilde{W}_{i}(K)} \biggl(\frac {\widetilde{W}_{j}(\Psi_{p}K)}{\widetilde{W}_{k}(\Psi_{p}K)} \biggr) ^{1/(k-j)}+ \frac{1}{\widetilde{W}_{i}(L)} \biggl(\frac{\widetilde {W}_{j}(\Psi_{p}L)}{\widetilde{W}_{k}(\Psi_{p}L)} \biggr)^{1/(k-j)}, \end{aligned}$$
(5.4)

with equality if and only if \(\Psi_{p}K\) and \(\Psi_{p}L\) are dilates.

Remark 5.4

Taking \(k=n\) in (5.4) and noting that \(\widetilde{W}_{n}(K)=\int_{S^{n-1}}\,dS(u)=n\omega_{n}\), (5.4) becomes the following inequality: If \(K,L\in{\mathcal {S}}^{n}\), \(0\leq i< n\), \(p< i-1\) and \(j\leq n-1\), then

$$\begin{aligned} \frac{\widetilde{W}_{j}(\Psi_{p}(K\breve {+}_{p}L))^{1/(n-j)}}{\widetilde{W}_{i}(K\breve{+}_{p} L)}\leq\frac{\widetilde{W}_{j}(\Psi_{p}K)^{1/(n-j)}}{\widetilde {W}_{i}(K)}+\frac{\widetilde{W}_{j}(\Psi_{p}L)^{1/(n-j)}}{\widetilde {W}_{i}(L)}, \end{aligned}$$
(5.5)

with equality if and only if \(\Psi_{p}K\) and \(\Psi_{p}L\) are dilates. Taking \(i=0\), \(j=0\) and \(k=n\) in (5.4), we have: If \(K,L\in{\mathcal {S}}^{n}\) and \(p<-1\), then

$$\begin{aligned} \frac{V(\Psi_{p}(K\breve{+}_{p}L))^{1/n}}{V(K\breve{+}_{p} L)}\leq\frac{V(\Psi_{p}K)^{1/n}}{V(K)}+\frac{V(\Psi _{p}L)^{1/n}}{V(L)}, \end{aligned}$$
(5.6)

with equality if and only if \(\Psi_{p}K\) and \(\Psi_{p}L\) are dilates.

6 Dresher’s inequalities for p-radial Blaschke and harmonic Blaschke additions

An extension of Beckenbach’s inequality (see [3], p. 27) was obtained by Dresher [25] by means of moment-space techniques.

Lemma 6.1

(Dresher’s inequality)

If \(p\geq1\geq r\geq0\), \(f,g\geq0\) and ϕ is a distribution function, then

$$\begin{aligned} \biggl(\frac{\int(f+g)^{p}\,d\phi}{\int(f+g)^{r}\,d\phi} \biggr)^{1/(p-r)}\leq \biggl(\frac{\int f^{p}\,d\phi}{\int f^{r}\,d\phi} \biggr) ^{1/(p-r)}+ \biggl(\frac{\int g^{p}\,d\phi}{\int g^{r}\,d\phi} \biggr)^{1/(p-r)}, \end{aligned}$$
(6.1)

with equality if and only if the functions f and g are proportional.

We are now in a position to prove Theorem 5.1. The following statement is just a slight reformulation of it.

Theorem 6.2

Let \(K,L\in{\mathcal {S}}^{n}\). If \(0\leq p< n-1\) and \(s,t\in{\Bbb {R}}\) satisfy \(s\geq1\geq t\geq0\), then

$$\begin{aligned} \biggl(\frac{\widetilde{W}_{n-s}(\Psi_{p}(K\widehat {+}_{p}L))}{\widetilde{W}_{n-t}(\Psi_{p}(K\widehat{+}_{p}L))} \biggr)^{1/(s-t)} \leq \biggl(\frac{\widetilde{W}_{n-s}(\Psi_{p}K)}{\widetilde {W}_{n-t}(\Psi_{p}K)} \biggr) ^{1/(s-t)}+ \biggl(\frac{\widetilde{W}_{n-s}(\Psi_{p}L)}{\widetilde {W}_{n-t}(\Psi_{p}L)} \biggr)^{1/(s-t)}, \end{aligned}$$
(6.2)

with equality if and only if \(\Psi_{p} K\) and \(\Psi_{p} L\) are dilates.

Proof

From (2.4), we obtain

$$\rho(K\widehat{+}_{p} L,\cdot)^{n-p-1}\ast\mu=\rho(K, \cdot)^{n-p-1}\ast\mu+\rho (L,\cdot)^{n-p-1}\ast\mu, $$

where μ is the generating measure of Ψ from Lemma 4.3. Hence, from (4.6), we obtain

$$\rho\bigl(\Psi_{p}(K\widehat{+}_{p}L),\cdot\bigr)=\rho( \Psi_{p}K,\cdot )+\rho(\Psi_{p}L,\cdot). $$

Therefore, by (1.6), we have

$$\begin{aligned} \widetilde{W}_{n-s}\bigl(\Psi_{p}(K\widehat{+}_{p}L) \bigr)=\frac{1}{n} \int _{S^{n-1}} \bigl(\rho(\Psi_{p}K,u)+\rho( \Psi_{p}L,u) \bigr)^{s}\,dS(u) \end{aligned}$$
(6.3)

and

$$\begin{aligned} \widetilde{W}_{n-t}\bigl(\Psi_{p}(K\widehat{+}_{p}L) \bigr)=\frac{1}{n} \int _{S^{n-1}} \bigl(\rho(\Psi_{p}K,u)+\rho( \Psi_{p}L,u) \bigr)^{t}\,dS(u). \end{aligned}$$
(6.4)

From (6.3), (6.4) and Lemma 6.1, we obtain

$$\begin{aligned} &\biggl(\frac{\widetilde{W}_{n-s}(\Psi_{p}(K\widehat {+}_{p}L))}{\widetilde{W}_{n-t}(\Psi_{p}(K\widehat{+}_{p}L))} \biggr)^{1/(s-t)} \\ &\quad= \biggl(\frac{\int_{S^{n-1}} (\rho(\Psi_{p}K,u)+\rho(\Psi _{p}L,u) )^{s}\,dS(u)}{ \int_{S^{n-1}} (\rho(\Psi_{p}K,u)+\rho(\Psi_{p}L,u) )^{t}\,dS(u)} \biggr)^{1/(s-t)} \\ &\quad\leq \biggl(\frac{\int_{S^{n-1}}\rho (\Psi_{p}K,u)^{s}\,dS(u)}{ \int_{S^{n-1}}\rho(\Psi_{p}K,u)^{t}\,dS(u)} \biggr)^{1/(s-t)}+ \biggl(\frac {\int_{S^{n-1}}\rho(\Psi_{p}L,u)^{s}\,dS(u)}{\int_{S^{n-1}}\rho(\Psi _{p}L,u)^{t}\,dS(u)} \biggr)^{1/(s-t)} \\ &\quad= \biggl(\frac{\widetilde{W}_{n-s}(\Psi_{p}K)}{\widetilde {W}_{n-t}(\Psi_{p}K)} \biggr) ^{1/(s-t)}+ \biggl(\frac{\widetilde{W}_{n-s}(\Psi_{p}L)}{\widetilde {W}_{n-t}(\Psi_{p}L)} \biggr)^{1/(s-t)}. \end{aligned}$$

Equality holds if and only if the functions \(\rho(\Psi_{p}K,u)\) and \(\rho(\Psi_{p}L,u)\) are proportional.

Taking \(s=n-i\) and \(t=n-j\) in Theorem 6.2, Theorem 6.2 becomes Theorem 5.1 stated in Section 5. If \(\Psi:\underbrace{{\mathcal {S}}^{n}\times\cdots\times{\mathcal {S}}^{n}}_{n-1}\rightarrow {\mathcal {S}}^{n}\) is the mixed intersection operator \({\bf I}:\underbrace{{\mathcal {S}}^{n}\times\cdots\times{\mathcal {S}}^{n}}_{n-1}\rightarrow{\mathcal {S}}^{n}\) in (6.2) and \(n-s=i\) and \(n-t=j\), we obtain the following result: If \(K,L\in{\mathcal {S}}^{n}\), \(0\leq p< n-1\) and \(i\leq n-1\leq j\leq n\), then

$$\begin{aligned} \biggl(\frac{\widetilde{W}_{i}(\mathbf{I}_{p}(K\widehat {+}_{p}L))}{\widetilde{W}_{j}(\mathbf{I}_{p}(K\widehat{+}_{p}L))} \biggr)^{1/(j-i)} \leq \biggl(\frac{\widetilde{W}_{i}({\bf I}_{p}K)}{\widetilde{W}_{j}(\mathbf{I}_{p}K)} \biggr) ^{1/(j-i)}+ \biggl(\frac{\widetilde{W}_{i}({\bf I}_{p}L)}{\widetilde{W}_{j}({\bf I}_{p}L)} \biggr)^{1/(j-i)}, \end{aligned}$$
(6.5)

with equality if and only if \(\mathbf{I}_{p}K\) and \(\mathbf{I}_{p}L\) are dilates. Taking \(j=n\) in (6.5) and noting that \(\widetilde{W}_{n}(K)=\int_{S^{n-1}}\,dS(u)=n\omega_{n}\), (6.5) becomes the following inequality: If \(K,L\in{\mathcal {S}}^{n}\), \(0\leq p< n-1\) and \(i\leq n-1\), then

$$\widetilde{W}_{i}\bigl(\mathbf{I}_{p}(K \widehat{+}_{p}L)\bigr)^{1/(n-i)} \leq\widetilde{W}_{i}({ \bf I}_{p}K)^{1/(n-i)}+\widetilde{W}_{i}( \mathbf{I}_{p}L)^{1/(n-i)}, $$

with equality if and only if \(\mathbf{I}_{p}K\) and \(\mathbf{I}_{p}L\) are dilates. □

We are now in a position to prove Theorem 5.3. The following statement is just a slight reformulation of it.

Theorem 6.3

Let \(K,L\in{\mathcal {S}}^{n}\). If \(0\leq i< n \), \(p< i-1\) and \(s,t\in{\Bbb {R}}\) satisfy \(s\geq1\geq t\geq0\), then

$$\begin{aligned} &\frac{1}{\widetilde{W}_{i}(K\breve{+}_{p} L)} \biggl(\frac{\widetilde{W}_{n-s}(\Psi_{p}(K\breve {+}_{p}L))}{\widetilde{W}_{n-t}(\Psi_{p}(K\breve{+}_{p}L))} \biggr)^{1/(s-t)} \\ &\quad\leq\frac{1}{\widetilde{W}_{i}(K)} \biggl(\frac {\widetilde{W}_{n-s}(\Psi_{p}K)}{\widetilde{W}_{n-t}(\Psi _{p}K)} \biggr) ^{1/(s-t)}+ \frac{1}{\widetilde{W}_{i}(L)} \biggl(\frac{\widetilde {W}_{n-s}(\Psi_{p}L)}{\widetilde{W}_{n-t}(\Psi_{p}L)} \biggr)^{1/(s-t)}, \end{aligned}$$
(6.6)

with equality if and only if \(\Psi_{p}K\) and \(\Psi_{p}L\) are dilates.

Proof

From (3.3), we obtain

$$\frac{\rho(K\breve{+}_{p} L,\cdot)^{n-p-1}\ast\mu}{\widetilde{W}_{i}(K\breve{+}_{p} L)}=\frac{\rho(K,\cdot)^{n-p-1}\ast\mu}{\widetilde {W}_{i}(K)}+\frac{\rho(L,\cdot)^{n-p-1}\ast\mu}{\widetilde{W}_{i}(L)}. $$

Hence, from (4.6), we obtain

$$\frac{\rho(\Psi_{p}(K\breve{+}_{p}L),\cdot)}{\widetilde {W}_{i}(K\breve{+}_{p} L)}=\frac{\rho(\Psi_{p}K,\cdot)}{\widetilde{W}_{i}(K)}+\frac{\rho (\Psi_{p}L,\cdot)}{\widetilde{W}_{i}(L)}. $$

By (1.6), we have

$$\begin{aligned} \frac{\widetilde{W}_{n-s}(\Psi_{p}(K\breve{+}_{p}L))}{\widetilde {W}_{i}(K\breve{+}_{p} L)^{s}}=\frac{1}{n} \int_{S^{n-1}} \biggl(\frac{\rho(\Psi _{p}K,u)}{\widetilde{W}_{i}(K)}+\frac{\rho(\Psi_{p}L,u)}{\widetilde {W}_{i}(L)} \biggr)^{s}\,dS(u) \end{aligned}$$
(6.7)

and

$$\begin{aligned} \frac{\widetilde{W}_{n-t}(\Psi_{p}(K\breve{+}_{p}L))}{\widetilde {W}_{i}(K\breve{+}_{p} L)^{t}}=\frac{1}{n} \int_{S^{n-1}} \biggl(\frac{\rho(\Psi _{p}K,u)}{\widetilde{W}_{i}(K)}+\frac{\rho(\Psi_{p}L,u)}{\widetilde {W}_{i}(L)} \biggr)^{t}\,dS(u). \end{aligned}$$
(6.8)

From (6.7), (6.8) and Lemma 6.1, we obtain

$$\begin{aligned} &\frac{1}{\widetilde{W}_{i}(K\breve{+}_{p} L)} \biggl(\frac{\widetilde{W}_{n-s}(\Psi_{p}(K\breve {+}_{p}L))}{\widetilde{W}_{n-t}(\Psi_{p}(K\breve{+}_{p}L))} \biggr)^{1/(s-t)} \\ &\quad= \biggl( \frac{\int_{S^{n-1}} (\frac{\rho(\Psi _{p}K,u)}{\widetilde{W}_{i}(K)}+\frac{\rho(\Psi_{p}L,u)}{\widetilde {W}_{i}(L)} )^{s}\,dS(u)}{ \int_{S^{n-1}} (\frac{\rho(\Psi_{p}K,u)}{\widetilde {W}_{i}(K)}+\frac{\rho(\Psi_{p}L,u)}{\widetilde{W}_{i}(L)} )^{t}\,dS(u)} \biggr)^{1/(s-t)} \\ &\quad\leq \biggl(\frac{\int_{S^{n-1}} (\frac{\rho(\Psi_{p}K,u)}{\widetilde {W}_{i}(K)} )^{s}\,dS(u)}{ \int_{S^{n-1}} (\frac{\rho(\Psi_{p}K,u)}{\widetilde {W}_{i}(K)} )^{t}\,dS(u)} \biggr)^{1/(s-t)}+ \biggl(\frac{\int _{S^{n-1}} (\frac{\rho(\Psi_{p}L,u)}{\widetilde {W}_{i}(L)} )^{s}\,dS(u)}{ \int_{S^{n-1}} (\frac{\rho(\Psi_{p}L,u)}{\widetilde {W}_{i}(L)} )^{t}\,dS(u)} \biggr)^{1/(s-t)} \\ &\quad=\frac{1}{\widetilde{W}_{i}(K)} \biggl(\frac {\widetilde{W}_{n-s}(\Psi_{p}K)}{\widetilde{W}_{n-t}(\Psi _{p}K)} \biggr) ^{1/(s-t)}+ \frac{1}{\widetilde{W}_{i}(L)} \biggl(\frac{\widetilde {W}_{n-s}(\Psi_{p}L)}{\widetilde{W}_{n-t}(\Psi_{p}L)} \biggr)^{1/(s-t)}, \end{aligned}$$

with equality if and only if \(\Psi_{p}K\) and \(\Psi_{p}L\) are dilates.

Taking \(s=n-j\) and \(t=n-k\) in Theorem 6.3, Theorem 6.3 becomes Theorem 5.3 stated in Section 5. If \(\Psi:\underbrace{{\mathcal {S}}^{n}\times\cdots\times{\mathcal {S}}^{n}}_{n-1}\rightarrow {\mathcal {S}}^{n}\) is the mixed intersection operator \({\bf I}:\underbrace{{\mathcal {S}}^{n}\times\cdots\times{\mathcal {S}}^{n}}_{n-1}\rightarrow{\mathcal {S}}^{n}\) in (6.6) and \(j=n-s\) and \(k=n-t\), we obtain the following result: If \(K,L\in{\mathcal {S}}^{n}\), \(0\leq i< n\), \(p< i-1\) and \(j\leq n-1\leq k\leq n\), then

$$\begin{aligned} &\frac{1}{\widetilde{W}_{i}(K\breve{+}_{p} L)} \biggl(\frac{\widetilde{W}_{j}({\bf I}_{p}(K\breve{+}_{p}L))}{\widetilde{W}_{k}({\bf I}_{p}(K\breve{+}_{p}L))} \biggr)^{1/(k-j)} \\ &\quad \leq \frac{1}{\widetilde{W}_{i}(K)} \biggl(\frac{\widetilde {W}_{j}({\bf I}_{p}K)}{\widetilde{W}_{k}(\mathbf{I}_{p}K)} \biggr) ^{1/(k-j)}+ \frac{1}{\widetilde{W}_{i}(L)} \biggl(\frac{\widetilde {W}_{j}({\bf I}_{p}L)}{\widetilde{W}_{k}({\bf I}_{p}L)} \biggr)^{1/(k-j)}, \end{aligned}$$
(6.9)

with equality if and only if \(\mathbf{I}_{p}K\) and \(\mathbf{I}_{p}L\) are dilates. Taking \(k=n\) in (6.9) and noting that \(\widetilde{W}_{n}(K)=\int_{S^{n-1}}\,dS(u)=n\omega_{n}\), (6.9) becomes the following inequality: If \(K,L\in{\mathcal {S}}^{n}\), \(0\leq i< n\), \(p< i-1\) and \(j\leq n-1\), then

$$\begin{aligned} \frac{\widetilde{W}_{j}({\bf I}_{p}(K\breve{+}_{p}L))^{1/(n-j)}}{\widetilde{W}_{i}(K\breve{+}_{p} L)}\leq\frac{\widetilde{W}_{j}({\bf I}_{p}K)^{1/(n-j)}}{\widetilde{W}_{i}(K)}+\frac{\widetilde {W}_{j}({\bf I}_{p}L)^{1/(n-j)}}{\widetilde{W}_{i}(L)}, \end{aligned}$$
(6.10)

with equality if and only if \(\mathbf{I}_{p}K\) and \(\mathbf{I}_{p}L\) are dilates. □

7 Conclusions

In the present study, we first revised and improved the concepts of radial Blaschke addition and harmonic Blaschke addition in an \(L_{p}\) space. Following this, we established Dresher’s inequalities (Brunn-Minkowski type) for the radial Blaschke-Minkowski homomorphisms with respect to the p-radial addition and the p-harmonic Blaschke addition.