Access the full text.
Sign up today, get DeepDyve free for 14 days.
References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.
1IntroductionLet Kn{{\mathcal{K}}}^{n}denote the set of convex bodies (compact, convex subsets with nonempty interiors) in the Euclidean nn-space Rn{{\mathbb{R}}}^{n}. For the set of convex bodies containing the origin in their interiors and the set of convex bodies whose centroid lie at the origin, we write Kon{{\mathcal{K}}}_{o}^{n}and Kcn{{\mathcal{K}}}_{c}^{n}, respectively. For the set of star bodies (about the origin) in Rn{{\mathbb{R}}}^{n}, we write Son{{\mathcal{S}}}_{o}^{n}. Let Sn−1{S}^{n-1}denote the unit sphere in Rn{{\mathbb{R}}}^{n}and V(K)V\left(K)denote the nn-dimensional volume of a body KK. For the standard unit ball BBin Rn{{\mathbb{R}}}^{n}, its volume is written by ωn=V(B){\omega }_{n}=V\left(B).If K∈KnK\in {{\mathcal{K}}}^{n}, then the support function of KK, hK=h(K,⋅):Rn→R,{h}_{K}=h\left(K,\cdot ):{{\mathbb{R}}}^{n}\to {\mathbb{R}},is defined by [3]h(K,x)=max{x⋅y:y∈K},x∈Rn,h\left(K,x)=\max \left\{x\cdot y:y\in K\right\},\hspace{1em}x\in {{\mathbb{R}}}^{n},where x⋅yx\cdot ydenotes the standard inner product of xxand yyin Rn{{\mathbb{R}}}^{n}.Let KKbe a compact star shaped (about the origin) in Rn{{\mathbb{R}}}^{n}, the radial function of KK, ρK=ρ(K,⋅):Rn⧹{0}→[0,+∞){\rho }_{K}=\rho \left(K,\cdot ):{{\mathbb{R}}}^{n}\setminus \left\{0\right\}\to {[}0,+\infty ), is defined by [4]ρ(K,x)=max{λ≥0:λx∈K},x∈Rn⧹{0}.\rho \left(K,x)=\max \left\{\lambda \ge 0:\lambda x\in K\right\},\hspace{1em}x\in {{\mathbb{R}}}^{n}\setminus \left\{0\right\}.If ρK{\rho }_{K}is positive and continuous, KKwill be called a star body (with respect to the origin). Two star bodies KKand LLare dilated (of one another) if ρK(u)/ρL(u){\rho }_{K}\left(u)\hspace{0.1em}\text{/}\hspace{0.1em}{\rho }_{L}\left(u)is independent of u∈Sn−1u\in {S}^{n-1}.For p≥1p\ge 1, the Lp{L}_{p}mixed volume, Vp(K,L){V}_{p}\left(K,L), of K,L∈KonK,L\in {{\mathcal{K}}}_{o}^{n}is defined by [5](1.1)Vp(K,L)=1n∫Sn−1hLp(v)dSp(K,v),{V}_{p}\left(K,L)=\frac{1}{n}\mathop{\int }\limits_{{S}^{n-1}}{h}_{L}^{p}\left(v){\rm{d}}{S}_{p}\left(K,v),where Sp(K,⋅){S}_{p}\left(K,\cdot )is the Lp{L}_{p}surface area measure.Based on the Lp{L}_{p}mixed volume (1.1), Lutwak [6] in 1996 introduced the notion of Lp{L}_{p}geominimal surface areas: For p≥1p\ge 1and K∈KonK\in {{\mathcal{K}}}_{o}^{n}, the Lp{L}_{p}geominimal surface area, Gp(K){G}_{p}\left(K), is defined by ωnpnGp(K)=inf{nVp(K,Q)V(Q∗)pn:Q∈Kon},{\omega }_{n}^{\frac{p}{n}}{G}_{p}\left(K)={\rm{\inf }}\left\{n{V}_{p}\left(K,Q)V{\left({Q}^{\ast })}^{\tfrac{p}{n}}:Q\in {{\mathcal{K}}}_{o}^{n}\right\},where Q∗{Q}^{\ast }denotes the polar of QQ. When p=1p=1, G1(K){G}_{1}\left(K)is just Petty’s classical geominimal surface area [7]. In particular, Lutwak [6] proved the following inequalities.Theorem 1.AIf K∈KonK\in {{\mathcal{K}}}_{o}^{n}and 1≤p<q1\le p\lt q, thenGp(K)nnnV(K)n−p1p≤Gq(K)nnnV(K)n−q1q,{\left(\frac{{G}_{p}{\left(K)}^{n}}{{n}^{n}V{\left(K)}^{n-p}}\right)}^{\tfrac{1}{p}}\le {\left(\frac{{G}_{q}{\left(K)}^{n}}{{n}^{n}V{\left(K)}^{n-q}}\right)}^{\tfrac{1}{q}},with equality if and only if K is p-self-minimal.Theorem 1.BIf K∈KonK\in {{\mathcal{K}}}_{o}^{n}and p≥1p\ge 1, thenωnGp(K)nnnV(K)n−p1p≤V(K)V(K∗),\hspace{-16em}{\omega }_{n}{\left(\frac{{G}_{p}{\left(K)}^{n}}{{n}^{n}V{\left(K)}^{n-p}}\right)}^{\tfrac{1}{p}}\le V\left(K)V\left({K}^{\ast }),with equality if and only if K is p-self-minimal.Very recently, Lutwak et al. [1] introduced the Lp{L}_{p}dual curvature measures as follows: Suppose p,q∈Rp,q\in {\mathbb{R}}. If K∈KonK\in {{\mathcal{K}}}_{o}^{n}while L∈SonL\in {{\mathcal{S}}}_{o}^{n}, then the Lp{L}_{p}dual curvature measures, C˜p,q(K,L,⋅){\widetilde{C}}_{p,q}\left(K,L,\cdot ), on Sn−1{S}^{n-1}are defined by ∫Sn−1g(v)dC˜p,q(K,L,v)=1n∫Sn−1g(αK(u))hK(αK(u))−pρK(u)qρL(u)n−qdu,\mathop{\int }\limits_{{S}^{n-1}}g\left(v){\rm{d}}{\widetilde{C}}_{p,q}\left(K,L,v)=\frac{1}{n}\mathop{\int }\limits_{{S}^{n-1}}g\left({\alpha }_{K}\left(u)){h}_{K}{\left({\alpha }_{K}\left(u))}^{-p}{\rho }_{K}{\left(u)}^{q}{\rho }_{L}{\left(u)}^{n-q}{\rm{d}}u,for each continuous g:Sn−1→Rg:{S}^{n-1}\to {\mathbb{R}}, where αK{\alpha }_{K}is the radial Gauss map (see [1], Section 3).Using the Lp{L}_{p}dual curvature measures, Lutwak et al. [1] defined the (p,q)\left(p,q)-mixed volumes as follows: For p,q∈Rp,q\in {\mathbb{R}}, K,Q∈KonK,Q\in {{\mathcal{K}}}_{o}^{n}and L∈SonL\in {{\mathcal{S}}}_{o}^{n}, the (p,q)\left(p,q)-mixed volume, V˜p,q(K,Q,L){\widetilde{V}}_{p,q}\left(K,Q,L), is defined by V˜p,q(K,Q,L)=∫Sn−1hQp(v)dC˜p,q(K,L,v).{\widetilde{V}}_{p,q}\left(K,Q,L)=\mathop{\int }\limits_{{S}^{n-1}}{h}_{Q}^{p}\left(v){\rm{d}}{\widetilde{C}}_{p,q}\left(K,L,v).Meanwhile, Lutwak et al. [1] gave the following formula of (p,q)\left(p,q)-mixed volume: (1.2)V˜p,q(K,Q,L)=1n∫Sn−1hQhK(αK(u))pρK(u)qρL(u)n−qdu.{\widetilde{V}}_{p,q}\left(K,Q,L)=\frac{1}{n}\mathop{\int }\limits_{{S}^{n-1}}\left(\frac{{h}_{Q}}{{h}_{K}}\right){\left({\alpha }_{K}\left(u))}^{p}{\rho }_{K}{\left(u)}^{q}{\rho }_{L}{\left(u)}^{n-q}{\rm{d}}u.In [1], Lutwak et al. introduced the Lp{L}_{p}mixed volume Vp(K,Q){V}_{p}\left(K,Q)of K,Q∈KonK,Q\in {{\mathcal{K}}}_{o}^{n}for all p∈Rp\in {\mathbb{R}}by Vp(K,Q)=1n∫Sn−1hQp(v)dSp(K,v).{V}_{p}\left(K,Q)=\frac{1}{n}\mathop{\int }\limits_{{S}^{n-1}}{h}_{Q}^{p}\left(v){\rm{d}}{S}_{p}\left(K,v).The case p≥1p\ge 1is just Lutwak’s Lp{L}_{p}mixed volume (1.1). Moreover, for q∈Rq\in {\mathbb{R}}and K,Q∈SonK,Q\in {{\mathcal{S}}}_{o}^{n}, they in [1] also defined the qth dual mixed volume, V˜q(K,Q){\widetilde{V}}_{q}\left(K,Q), by V˜q(K,Q)=1n∫Sn−1ρKq(v)ρQn−q(v)dv.{\widetilde{V}}_{q}\left(K,Q)=\frac{1}{n}\mathop{\int }\limits_{{S}^{n-1}}{\rho }_{K}^{q}\left(v){\rho }_{Q}^{n-q}\left(v){\rm{d}}v.By (1.2), the following special cases are showed: (1.3)V˜p,q(K,Q,K)=Vp(K,Q),\hspace{-13.75em}{\widetilde{V}}_{p,q}\left(K,Q,K)={V}_{p}\left(K,Q),(1.4)V˜p,q(K,K,L)=V˜q(K,L),\hspace{-13.75em}{\widetilde{V}}_{p,q}\left(K,K,L)={\widetilde{V}}_{q}\left(K,L),(1.5)V˜p,n(K,Q,L)=Vp(K,Q).\hspace{-13.75em}{\widetilde{V}}_{p,n}\left(K,Q,L)={V}_{p}\left(K,Q).The (p,q)\left(p,q)-mixed volumes unify the mixed volumes of convex bodies in the Lp{L}_{p}Brunn-Minkowski theory and the dual mixed volumes of star bodies in the Lp{L}_{p}dual Brunn-Minkowski theory. In the last 20 years, the Lp{L}_{p}Brunn-Minkowski theory and its dual theory have been developed very rapidly, see e.g., [3, 4,5,6,8, 9,10,11, 12,13,14, 15,16,17, 18,19,20, 21,22,23].Based on the (p,q)\left(p,q)-mixed volumes, Feng and He in [2] introduced the concept of (p,q)\left(p,q)-mixed geominimal surface areas as follows:Definition 1.1For p,q∈Rp,q\in {\mathbb{R}}, K∈KonK\in {{\mathcal{K}}}_{o}^{n}and L∈SonL\in {{\mathcal{S}}}_{o}^{n}, the (p,q)\left(p,q)-mixed geominimal surface areas, G˜p,q(K,L){\widetilde{G}}_{p,q}\left(K,L), of KKand LLare defined by (1.6)ωnpnG˜p,q(K,L)=inf{nV˜p,q(K,Q,L)V(Q∗)pn:Q∈Kon}.\hspace{-25.45em}{\omega }_{n}^{\frac{p}{n}}{\widetilde{G}}_{p,q}\left(K,L)={\rm{\inf }}\left\{n{\widetilde{V}}_{p,q}\left(K,Q,L)V{\left({Q}^{\ast })}^{\tfrac{p}{n}}:Q\in {{\mathcal{K}}}_{o}^{n}\right\}.If L=KL=Kor q=nq=nin (1.6), then from (1.3) or (1.5) we see that for p≥1p\ge 1the definition is just Lutwak’s Lp{L}_{p}geominimal surface area in [6]. For the studies of Lp{L}_{p}geominimal surface areas, some results have been obtained in these articles (see, e.g., [24,25,26, 27,28,29, 30,31,32, 33,34,35]).According to the definition of (p,q)\left(p,q)-mixed geominimal surface areas, Feng and He [2] extended Theorem 1.A to the following result:Theorem 1.CSuppose q∈Rq\in {\mathbb{R}}. If K∈KonK\in {{\mathcal{K}}}_{o}^{n}and L∈SonL\in {{\mathcal{S}}}_{o}^{n}, then for 0<r<s0\lt r\lt s, G˜r,q(K,L)nnnV˜q(K,L)n−r1r≤G˜s,q(K,L)nnnV˜q(K,L)n−s1s,{\left(\frac{{\widetilde{G}}_{r,q}{\left(K,L)}^{n}}{{n}^{n}{\widetilde{V}}_{q}{\left(K,L)}^{n-r}}\right)}^{\tfrac{1}{r}}\le {\left(\frac{{\widetilde{G}}_{s,q}{\left(K,L)}^{n}}{{n}^{n}{\widetilde{V}}_{q}{\left(K,L)}^{n-s}}\right)}^{\tfrac{1}{s}},with equality if and only if K and L are dilates.Obviously, Theorem 1.C for L=KL=Kand 1≤r<s1\le r\lt simplies Theorem 1.A.In addition, they [2] gave a Brunn-Minkowski-type inequality for the (p,q)\left(p,q)-mixed geominimal surface areas as follows:Theorem 1.DSuppose p,q∈Rp,q\in {\mathbb{R}}such that 0<n−qq<10\lt \frac{n-q}{q}\lt 1and q≠nq\ne n, and let λ,μ∈R\lambda ,\mu \in {\mathbb{R}}. If K∈KonK\in {{\mathcal{K}}}_{o}^{n}, and L1,L2∈Son{L}_{1},{L}_{2}\in {{\mathcal{S}}}_{o}^{n}, thenG˜p,q(K,λ⋅L1+˜qμ⋅L2)qn−q≥λG˜p,q(K,L1)qn−q+μG˜p,q(K,L2)qn−q,{\widetilde{G}}_{p,q}{\left(K,\lambda \cdot {L}_{1}{\widetilde{+}}_{q}\mu \cdot {L}_{2})}^{\tfrac{q}{n-q}}\ge \lambda {\widetilde{G}}_{p,q}{\left(K,{L}_{1})}^{\tfrac{q}{n-q}}+\mu {\widetilde{G}}_{p,q}{\left(K,{L}_{2})}^{\tfrac{q}{n-q}},with equality if and only if L1{L}_{1}and L2{L}_{2}are dilates.2Main resultsHere, λ⋅L1+˜qμ⋅L2\lambda \cdot {L}_{1}{\widetilde{+}}_{q}\mu \cdot {L}_{2}is the Lq{L}_{q}-radial combination of L1{L}_{1}and L2{L}_{2}(see (2.1)).In this article, associated with the (p,q)\left(p,q)-mixed geominimal surface areas, we first establish two monotonic inequalities as follows:Theorem 1.1Suppose p,q∈Rp,q\in {\mathbb{R}}. If K∈KonK\in {{\mathcal{K}}}_{o}^{n}and L∈SonL\in {{\mathcal{S}}}_{o}^{n}, then for 1≤p<q≤n−11\le p\lt q\le n-1, (1.7)G˜p,n−p(K,L)nnnV(K)n−p1p≤G˜q,n−q(K,L)nnnV(K)n−q1q,{\left(\frac{{\widetilde{G}}_{p,n-p}{\left(K,L)}^{n}}{{n}^{n}V{\left(K)}^{n-p}}\right)}^{\tfrac{1}{p}}\le {\left(\frac{{\widetilde{G}}_{q,n-q}{\left(K,L)}^{n}}{{n}^{n}V{\left(K)}^{n-q}}\right)}^{\tfrac{1}{q}},with equality if and only if K and L are dilates.Theorem 1.2Suppose p,q∈Rp,q\in {\mathbb{R}}such that 1≤p<q1\le p\lt q. If K∈KonK\in {{\mathcal{K}}}_{o}^{n}and L∈SonL\in {{\mathcal{S}}}_{o}^{n}, then(1.8)G˜p,p(K,L)nnnV(L)n−p1p≤G˜q,q(K,L)nnnV(L)n−q1q,{\left(\frac{{\widetilde{G}}_{p,p}{\left(K,L)}^{n}}{{n}^{n}V{\left(L)}^{n-p}}\right)}^{\tfrac{1}{p}}\le {\left(\frac{{\widetilde{G}}_{q,q}{\left(K,L)}^{n}}{{n}^{n}V{\left(L)}^{n-q}}\right)}^{\tfrac{1}{q}},with equality if and only if K and L are dilates.Remark 1.1When L=KL=Kand KKis pp-self-minimal, Theorems 1.1 and 1.2 both become Theorem 1.A.Next, we obtain an affine isoperimetric inequality for the (p,q)\left(p,q)-mixed geominimal surface areas.Theorem 1.3Suppose p,q∈Rp,q\in {\mathbb{R}}such that p>0p\gt 0and 0<q≤n0\lt q\le n. If K∈KonK\in {{\mathcal{K}}}_{o}^{n}and L∈SonL\in {{\mathcal{S}}}_{o}^{n}, then(1.9)ωnG˜p,q(K,L)nnnV(K)q−pV(L)n−q1p≤V(K)V(K∗),\hspace{-24.7em}{\omega }_{n}{\left(\frac{{\widetilde{G}}_{p,q}{\left(K,L)}^{n}}{{n}^{n}V{\left(K)}^{q-p}V{\left(L)}^{n-q}}\right)}^{\tfrac{1}{p}}\le V\left(K)V\left({K}^{\ast }),for 0<q<n0\lt q\lt nequality holds when K and L are dilates.Remark 1.2If q=nq=nand p≥1p\ge 1, Theorem 1.3 only contains Lutwak’s inequality.Finally, together with the Lq{L}_{q}harmonic Blaschke combination, we give the following Brunn-Minkowski-type inequality for the (p,q)\left(p,q)-mixed geominimal surface areas.Theorem 1.4Suppose p,q∈Rp,q\in {\mathbb{R}}such that 0<q<n0\lt q\lt n, and let λ,μ≥0\lambda ,\mu \ge 0(not both zero). If K∈KonK\in {{\mathcal{K}}}_{o}^{n}and L1,L2∈Son{L}_{1},{L}_{2}\in {{\mathcal{S}}}_{o}^{n}, then(1.10)G˜p,q(K,λ∗L1+^qμ∗L2)n+qn−qV(λ∗L1+^qμ∗L2)≥λG˜p,q(K,L1)n+qn−qV(L1)+μG˜p,q(K,L2)n+qn−qV(L2),\frac{{\widetilde{G}}_{p,q}{\left(K,\lambda \ast {L}_{1}{\widehat{+}}_{q}\mu \ast {L}_{2})}^{\tfrac{n+q}{n-q}}}{V\left(\lambda \ast {L}_{1}{\widehat{+}}_{q}\mu \ast {L}_{2})}\ge \lambda \frac{{\widetilde{G}}_{p,q}{\left(K,{L}_{1})}^{\tfrac{n+q}{n-q}}}{V\left({L}_{1})}+\mu \frac{{\widetilde{G}}_{p,q}{\left(K,{L}_{2})}^{\tfrac{n+q}{n-q}}}{V\left({L}_{2})},with equality if and only if L1{L}_{1}and L2{L}_{2}are dilates.Here, λ∗L1+^qμ∗L2\lambda \ast {L}_{1}{\widehat{+}}_{q}\mu \ast {L}_{2}is the Lq{L}_{q}harmonic Blaschke combination of L1,L2∈Son{L}_{1},{L}_{2}\in {{\mathcal{S}}}_{o}^{n}(see (2.2)).3BackgroundIn order to complete the proofs of Theorems 1.1–1.4, we collect some basic facts about convex bodies and star bodies.If EEis a nonempty subset in Rn{{\mathbb{R}}}^{n}, then the polar set, E∗{E}^{\ast }, of EEis defined by [1,2]E∗={x∈Rn:x⋅y≤1,y∈E}.{E}^{\ast }=\left\{x\in {{\mathbb{R}}}^{n}:x\cdot y\le 1,\hspace{1em}y\in E\right\}.Meanwhile, it is easy to get that (K∗)∗=K{\left({K}^{\ast })}^{\ast }=Kfor all K∈KonK\in {{\mathcal{K}}}_{o}^{n}.For K,L∈SonK,L\in {{\mathcal{S}}}_{o}^{n}, q≠0q\ne 0and λ,μ≥0\lambda ,\mu \ge 0(not both zero), the Lq{L}_{q}radial combination, λ⋅K+˜qμ⋅L∈Son\lambda \cdot K{\tilde{+}}_{q}\mu \cdot L\in {{\mathcal{S}}}_{o}^{n}, of KKand LLis defined by [1](2.1)ρ(λ⋅K+˜qμ⋅L,⋅)q=λρ(K,⋅)q+μρ(L,⋅)q,\hspace{-19.9em}\rho {\left(\lambda \cdot K{\tilde{+}}_{q}\mu \cdot L,\cdot )}^{q}=\lambda \rho {\left(K,\cdot )}^{q}+\mu \rho {\left(L,\cdot )}^{q},where the operation “+˜q{\tilde{+}}_{q}” is called Lq{L}_{q}radial addition, λ⋅K\lambda \cdot Kdenotes Lq{L}_{q}radial scalar multiplication and λ⋅K=λ1qK\lambda \cdot K={\lambda }^{\tfrac{1}{q}}K.For K,L∈SonK,L\in {{\mathcal{S}}}_{o}^{n}, q>0q\gt 0and λ,μ≥0\lambda ,\mu \ge 0(not both zero), the Lq{L}_{q}harmonic Blaschke combination, λ∗K+^qμ∗L∈Son\lambda \ast K{\widehat{+}}_{q}\mu \ast L\in {{\mathcal{S}}}_{o}^{n}, of KKand LLis defined by [36](2.2)ρ(λ∗K+^qμ∗L,⋅)n+qV(λ∗K+^qμ∗L)=λρ(K,⋅)n+qV(K)+μρ(L,⋅)n+qV(L),\hspace{-23.8em}\frac{\rho {\left(\lambda \ast K{\widehat{+}}_{q}\mu \ast L,\cdot )}^{n+q}}{V\left(\lambda \ast K{\widehat{+}}_{q}\mu \ast L)}=\lambda \frac{\rho {\left(K,\cdot )}^{n+q}}{V\left(K)}+\mu \frac{\rho {\left(L,\cdot )}^{n+q}}{V\left(L)},where the operation “+^q{\widehat{+}}_{q}” is called Lq{L}_{q}harmonic Blaschke addition, λ∗K\lambda \ast Kdenotes Lq{L}_{q}harmonic Blaschke scalar multiplication and λ∗K=λ1qK\lambda \ast K={\lambda }^{\tfrac{1}{q}}K. When λ=μ=1\lambda =\mu =1, K+^qLK{\widehat{+}}_{q}Lis called Lq{L}_{q}harmonic Blaschke sum.4Proofs of main theoremsIn this section, we will prove Theorems 1.1–1.4. To complete the proof of Theorem 1.1, we require the following lemma.Lemma 3.1[37] Suppose p,q∈Rp,q\in {\mathbb{R}}. If K,Q∈KonK,Q\in {{\mathcal{K}}}_{o}^{n}and L∈SonL\in {{\mathcal{S}}}_{o}^{n}, then for 0<p<q≤n−10\lt p\lt q\le n-1, (3.1)V˜p,n−p(K,Q,L)V(K)1p≤V˜q,n−q(K,Q,L)V(K)1q,{\left[\frac{{\widetilde{V}}_{p,n-p}\left(K,Q,L)}{V\left(K)}\right]}^{\tfrac{1}{p}}\le {\left[\frac{{\widetilde{V}}_{q,n-q}\left(K,Q,L)}{V\left(K)}\right]}^{\tfrac{1}{q}},with equality if and only if K and L are dilates.Proof of Theorem 1.1Together (1.6) with inequality (3.1), we obtain that for 1≤p<q≤n−11\le p\lt q\le n-1, ωnG˜p,n−p(K,L)nnnV(K)n−p1p=infV˜p,n−p(K,Q,L)V(K)npV(K)V(Q∗):Q∈Kon≤infV˜q,n−q(K,Q,L)V(K)nqV(K)V(Q∗):Q∈Kon=ωnG˜q,n−q(K,L)nnnV(K)n−q1q.\begin{array}{rcl}{\omega }_{n}{\left(\frac{{\widetilde{G}}_{p,n-p}{\left(K,L)}^{n}}{{n}^{n}V{\left(K)}^{n-p}}\right)}^{\tfrac{1}{p}}& =& \inf \left\{{\left(\frac{{\widetilde{V}}_{p,n-p}\left(K,Q,L)}{V\left(K)}\right)}^{\tfrac{n}{p}}V\left(K)V\left({Q}^{\ast }):Q\in {{\mathcal{K}}}_{o}^{n}\right\}\\ & \le & \inf \left\{{\left(\frac{{\widetilde{V}}_{q,n-q}\left(K,Q,L)}{V\left(K)}\right)}^{\tfrac{n}{q}}V\left(K)V\left({Q}^{\ast }):Q\in {{\mathcal{K}}}_{o}^{n}\right\}\\ & =& {\omega }_{n}{\left(\frac{{\widetilde{G}}_{q,n-q}{\left(K,L)}^{n}}{{n}^{n}V{\left(K)}^{n-q}}\right)}^{\tfrac{1}{q}}.\end{array}This is G˜p,n−p(K,L)nnnV(K)n−p1p≤G˜q,n−q(K,L)nnnV(K)n−q1q.{\left(\frac{{\widetilde{G}}_{p,n-p}{\left(K,L)}^{n}}{{n}^{n}V{\left(K)}^{n-p}}\right)}^{\tfrac{1}{p}}\le {\left(\frac{{\widetilde{G}}_{q,n-q}{\left(K,L)}^{n}}{{n}^{n}V{\left(K)}^{n-q}}\right)}^{\tfrac{1}{q}}.This yields inequality (1.7). According to the equality condition of inequality (3.1), we see that the equality of the above inequality holds if and only if KKand LLare dilates.□Lemma 3.2[37] Suppose p,q∈Rp,q\in {\mathbb{R}}satisfy 1≤p<q1\le p\lt q. If K,Q∈KonK,Q\in {{\mathcal{K}}}_{o}^{n}and L∈SonL\in {{\mathcal{S}}}_{o}^{n}, then(3.2)V˜p,p(K,Q,L)V(L)1p≤V˜q,q(K,Q,L)V(L)1q,{\left[\frac{{\widetilde{V}}_{p,p}\left(K,Q,L)}{V\left(L)}\right]}^{\tfrac{1}{p}}\le {\left[\frac{{\widetilde{V}}_{q,q}\left(K,Q,L)}{V\left(L)}\right]}^{\tfrac{1}{q}},with equality if and only if K and L are dilates.Proof of Theorem 1.2From Definition 1.1 and (3.2) we see that for 1≤p<q1\le p\lt q, ωnG˜p,p(K,L)nnnV(L)n−p1p=infV˜p,p(K,Q,L)V(L)npV(L)V(Q∗):Q∈Kon≤infV˜q,q(K,Q,L)V(L)nqV(L)V(Q∗):Q∈Kon=ωnG˜q,q(K,L)nnnV(L)n−q1q.\begin{array}{rcl}{\omega }_{n}{\left(\frac{{\widetilde{G}}_{p,p}{\left(K,L)}^{n}}{{n}^{n}V{\left(L)}^{n-p}}\right)}^{\tfrac{1}{p}}& =& \inf \left\{{\left(\frac{{\widetilde{V}}_{p,p}\left(K,Q,L)}{V\left(L)}\right)}^{\tfrac{n}{p}}V\left(L)V\left({Q}^{\ast }):Q\in {{\mathcal{K}}}_{o}^{n}\right\}\\ & \le & \inf \left\{{\left(\frac{{\widetilde{V}}_{q,q}\left(K,Q,L)}{V\left(L)}\right)}^{\tfrac{n}{q}}V\left(L)V\left({Q}^{\ast }):Q\in {{\mathcal{K}}}_{o}^{n}\right\}\\ & =& {\omega }_{n}{\left(\frac{{\widetilde{G}}_{q,q}{\left(K,L)}^{n}}{{n}^{n}V{\left(L)}^{n-q}}\right)}^{\tfrac{1}{q}}.\end{array}This is, for 1≤p<q1\le p\lt q, G˜p,p(K,L)nnnV(L)n−p1p≤G˜q,q(K,L)nnnV(L)n−q1q.{\left(\frac{{\widetilde{G}}_{p,p}{\left(K,L)}^{n}}{{n}^{n}V{\left(L)}^{n-p}}\right)}^{\tfrac{1}{p}}\le {\left(\frac{{\widetilde{G}}_{q,q}{\left(K,L)}^{n}}{{n}^{n}V{\left(L)}^{n-q}}\right)}^{\tfrac{1}{q}}.This gives inequality (1.8). From the equality condition of inequality (3.2), we see that the equality holds in (1.8) if and only if K,LK,Lare dilates.□Lemma 3.3[2] If K,L∈SonK,L\in {{\mathcal{S}}}_{o}^{n}, 0<q≤n0\lt q\le n, then(3.3)V˜q(K,L)≤V(K)qnV(L)n−qn.{\widetilde{V}}_{q}\left(K,L)\le V{\left(K)}^{\tfrac{q}{n}}V{\left(L)}^{\tfrac{n-q}{n}}.For 0<q<n0\lt q\lt n, equality holds in (3.3) if and only if K and L are dilates; for q=nq=n, (3.3) becomes an equality.Proof of Theorem 1.3Since p>0p\gt 0, it follows from (1.6) that (3.4)ωnG˜p,q(K,L)np=inf{nnpV˜p,q(K,Q,L)npV(Q∗):Q∈Kon}.{\omega }_{n}{\widetilde{G}}_{p,q}{\left(K,L)}^{\tfrac{n}{p}}={\rm{\inf }}\left\{{n}^{\tfrac{n}{p}}{\widetilde{V}}_{p,q}{\left(K,Q,L)}^{\tfrac{n}{p}}V\left({Q}^{\ast }):Q\in {{\mathcal{K}}}_{o}^{n}\right\}.Taking Q=KQ=Kin (3.4), it follows from (1.4) and (3.3) that for 0<q≤n0\lt q\le n, ωnG˜p,q(K,L)np≤nnpV˜p,q(K,K,L)npV(K∗)=nnpV˜q(K,L)npV(K∗)≤nnpV(K)qpV(L)n−qpV(K∗),{\omega }_{n}{\widetilde{G}}_{p,q}{\left(K,L)}^{\tfrac{n}{p}}\le {n}^{\tfrac{n}{p}}{\widetilde{V}}_{p,q}{\left(K,K,L)}^{\tfrac{n}{p}}V\left({K}^{\ast })={n}^{\tfrac{n}{p}}{\widetilde{V}}_{q}{\left(K,L)}^{\tfrac{n}{p}}V\left({K}^{\ast })\le {n}^{\tfrac{n}{p}}V{\left(K)}^{\tfrac{q}{p}}V{\left(L)}^{\tfrac{n-q}{p}}V\left({K}^{\ast }),i.e., ωnG˜p,q(K,L)nnnV(K)q−pV(L)n−q1p≤V(K)V(K∗).\hspace{-24.85em}{\omega }_{n}{\left(\frac{{\widetilde{G}}_{p,q}{\left(K,L)}^{n}}{{n}^{n}V{\left(K)}^{q-p}V{\left(L)}^{n-q}}\right)}^{\tfrac{1}{p}}\le V\left(K)V\left({K}^{\ast }).By the equality condition of inequality (3.3) and Definition 1.1, for 0<q<n0\lt q\lt nequality holds in (1.9) when KKand LLare dilates.□Lemma 3.4Suppose p,q∈Rp,q\in {\mathbb{R}}such that 0<q<n0\lt q\lt n, and let λ,μ>0\lambda ,\mu \gt 0. If K,Q∈KonK,Q\in {{\mathcal{K}}}_{o}^{n}and L1,L2∈Son{L}_{1},{L}_{2}\in {{\mathcal{S}}}_{o}^{n}, then(3.5)V˜p,q(K,Q,λ∗L1+^qμ∗L2)n+qn−qV(λ∗L1+^qμ∗L2)≥λV˜p,q(K,Q,L1)n+qn−qV(L1)+μV˜p,q(K,Q,L2)n+qn−qV(L2),\hspace{1em}\frac{{\widetilde{V}}_{p,q}{\left(K,Q,\lambda \ast {L}_{1}{\widehat{+}}_{q}\mu \ast {L}_{2})}^{\tfrac{n+q}{n-q}}}{V\left(\lambda \ast {L}_{1}{\widehat{+}}_{q}\mu \ast {L}_{2})}\ge \lambda \frac{{\widetilde{V}}_{p,q}{\left(K,Q,{L}_{1})}^{\tfrac{n+q}{n-q}}}{V\left({L}_{1})}+\mu \frac{{\widetilde{V}}_{p,q}{\left(K,Q,{L}_{2})}^{\tfrac{n+q}{n-q}}}{V\left({L}_{2})},with equality if and only if L1{L}_{1}and L2{L}_{2}are dilates.ProofSince 0<q<n0\lt q\lt n, thus 0<n−qn+q<10\lt \frac{n-q}{n+q}\lt 1. From (1.2), (2.2) and Minkowski’s integral inequality [38], we get that for any Q∈KonQ\in {{\mathcal{K}}}_{o}^{n}, V˜p,q(K,Q,λ∗L1+^qμ∗L2)n+qn−qV(λ∗L1+^qμ∗L2)=1n∫Sn−1hQhKp(αK(u))ρKq(u)ρλ∗L1+^qμ∗L2n−q(u)dun+qn−qV(λ∗L1+^qμ∗L2)=1n∫Sn−1hQhKp(n+q)n−q(αK(u))ρK(u)q(n+q)n−qρλ∗L1+^qμ∗L2n+q(u)V(λ∗L1+^qμ∗L2)n−qn+qdun+qn−q=1n∫Sn−1hQhKp(n+q)n−q(αK(u))ρK(u)q(n+q)n−qλρL1n+q(u)V(L1)+μρL2n+q(u)V(L2)n−qn+qdun+qn−q≥λ1n∫Sn−1hQhKp(αK(u))ρKq(u)ρL1n−q(u)dun+qn−qV(L1)+μ1n∫Sn−1hQhKp(αK(u))ρKq(u)ρL2n−q(u)dun+qn−qV(L2)=λV˜p,q(K,Q,L1)n+qn−qV(L1)+μV˜p,q(K,Q,L2)n+qn−qV(L2).\begin{array}{rcl}\frac{{\widetilde{V}}_{p,q}{\left(K,Q,\lambda \ast {L}_{1}{\widehat{+}}_{q}\mu \ast {L}_{2})}^{\tfrac{n+q}{n-q}}}{V\left(\lambda \ast {L}_{1}{\widehat{+}}_{q}\mu \ast {L}_{2})}& =& \frac{{\left[\frac{1}{n}{\displaystyle \int }_{{S}^{n-1}}{\left(\frac{{h}_{Q}}{{h}_{K}}\right)}^{p}\left({\alpha }_{K}\left(u)){\rho }_{K}^{q}\left(u){\rho }_{\lambda \ast {L}_{1}{\widehat{+}}_{q}\mu \ast {L}_{2}}^{n-q}\left(u){\rm{d}}u\right]}^{\tfrac{n+q}{n-q}}}{V\left(\lambda \ast {L}_{1}{\widehat{+}}_{q}\mu \ast {L}_{2})}\\ & =& {\left[\frac{1}{n}\mathop{\displaystyle \int }\limits_{{S}^{n-1}}{\left({\left(\frac{{h}_{Q}}{{h}_{K}}\right)}^{\frac{p\left(n+q)}{n-q}}\left({\alpha }_{K}\left(u)){\rho }_{K}{\left(u)}^{\frac{q\left(n+q)}{n-q}}\frac{{\rho }_{\lambda \ast {L}_{1}{\widehat{+}}_{q}\mu \ast {L}_{2}}^{n+q}\left(u)}{V\left(\lambda \ast {L}_{1}{\widehat{+}}_{q}\mu \ast {L}_{2})}\right)}^{\frac{n-q}{n+q}}{\rm{d}}u\right]}^{\tfrac{n+q}{n-q}}\\ & =& {\left[\frac{1}{n}\mathop{\displaystyle \int }\limits_{{S}^{n-1}}{\left({\left(\frac{{h}_{Q}}{{h}_{K}}\right)}^{\frac{p\left(n+q)}{n-q}}\left({\alpha }_{K}\left(u)){\rho }_{K}{\left(u)}^{\frac{q\left(n+q)}{n-q}}\left(\lambda \frac{{\rho }_{{L}_{1}}^{n+q}\left(u)}{V\left({L}_{1})}+\mu \frac{{\rho }_{{L}_{2}}^{n+q}\left(u)}{V\left({L}_{2})}\right)\right)}^{\frac{n-q}{n+q}}{\rm{d}}u\right]}^{\tfrac{n+q}{n-q}}\\ & \ge & \lambda \frac{{\left[\frac{1}{n}{\displaystyle \int }_{{S}^{n-1}}{\left(\frac{{h}_{Q}}{{h}_{K}}\right)}^{p}\left({\alpha }_{K}\left(u)){\rho }_{K}^{q}\left(u){\rho }_{{L}_{1}}^{n-q}\left(u){\rm{d}}u\right]}^{\tfrac{n+q}{n-q}}}{V\left({L}_{1})}\\ & & \phantom{\rule[-1.5em]{}{0ex}}+\mu \frac{{\left[\frac{1}{n}{\displaystyle \int }_{{S}^{n-1}}{\left(\frac{{h}_{Q}}{{h}_{K}}\right)}^{p}\left({\alpha }_{K}\left(u)){\rho }_{K}^{q}\left(u){\rho }_{{L}_{2}}^{n-q}\left(u){\rm{d}}u\right]}^{\tfrac{n+q}{n-q}}}{V\left({L}_{2})}\\ & =& \lambda \frac{{\widetilde{V}}_{p,q}{\left(K,Q,{L}_{1})}^{\tfrac{n+q}{n-q}}}{V\left({L}_{1})}+\mu \frac{{\widetilde{V}}_{p,q}{\left(K,Q,{L}_{2})}^{\tfrac{n+q}{n-q}}}{V\left({L}_{2})}.\end{array}This yields inequality (3.5).By the equality condition of Minkowski’s integral inequality, we see that equality holds in (3.5) if and only if L1{L}_{1}and L2{L}_{2}are dilates.□Proof of Theorem 1.4Because of 0<q<n0\lt q\lt n, thus n+qn−q>0\frac{n+q}{n-q}\gt 0. Hence, by (1.6) and (3.5) we have ωnpnG˜p,q(K,λ∗L1+^qμ∗L2)n+qn−qV(λ∗L1+^qμ∗L2)=inf[nV˜p,q(K,Q,λ∗L1+^qμ∗L2)]n+qn−qV(λ∗L1+^qμ∗L2)V(Q∗)pnn+qn−q:Q∈Kon≥λinf[nV˜p,q(K,Q,L1)]n+qn−qV(L1)V(Q∗)pnn+qn−q:Q∈Kon+μinf[nV˜p,q(K,Q,L2)]n+qn−qV(L2)V(Q∗)pnn+qn−q:Q∈Kon=λωnpnG˜p,q(K,L1)n+qn−qV(L1)+μωnpnG˜p,q(K,L2)n+qn−qV(L2),\begin{array}{rcl}\frac{{\left[{\omega }_{n}^{\frac{p}{n}}{\widetilde{G}}_{p,q}\left(K,\lambda \ast {L}_{1}{\widehat{+}}_{q}\mu \ast {L}_{2})\right]}^{\tfrac{n+q}{n-q}}}{V\left(\lambda \ast {L}_{1}{\widehat{+}}_{q}\mu \ast {L}_{2})}& =& \inf \left\{\frac{{\left[n{\widetilde{V}}_{p,q}\left(K,Q,\lambda \ast {L}_{1}{\widehat{+}}_{q}\mu \ast {L}_{2})]}^{\tfrac{n+q}{n-q}}}{V\left(\lambda \ast {L}_{1}{\widehat{+}}_{q}\mu \ast {L}_{2})}{\left[V{\left({Q}^{\ast })}^{\frac{p}{n}}\right]}^{\tfrac{n+q}{n-q}}:Q\in {{\mathcal{K}}}_{o}^{n}\right\}\\ & \ge & \lambda \inf \left\{\frac{{\left[n{\widetilde{V}}_{p,q}\left(K,Q,{L}_{1})]}^{\tfrac{n+q}{n-q}}}{V\left({L}_{1})}{\left[V{\left({Q}^{\ast })}^{\frac{p}{n}}\right]}^{\tfrac{n+q}{n-q}}:Q\in {{\mathcal{K}}}_{o}^{n}\right\}\\ & & +\mu \inf \left\{\frac{{\left[n{\widetilde{V}}_{p,q}\left(K,Q,{L}_{2})]}^{\tfrac{n+q}{n-q}}}{V\left({L}_{2})}{\left[V{\left({Q}^{\ast })}^{\frac{p}{n}}\right]}^{\tfrac{n+q}{n-q}}:Q\in {{\mathcal{K}}}_{o}^{n}\right\}\\ & =& \lambda \frac{{\left[{\omega }_{n}^{\frac{p}{n}}{\widetilde{G}}_{p,q}\left(K,{L}_{1})\right]}^{\tfrac{n+q}{n-q}}}{V\left({L}_{1})}+\mu \frac{{\left[{\omega }_{n}^{\frac{p}{n}}{\widetilde{G}}_{p,q}\left(K,{L}_{2})\right]}^{\tfrac{n+q}{n-q}}}{V\left({L}_{2})},\end{array}i.e., G˜p,q(K,λ∗L1+^qμ∗L2)n+qn−qV(λ∗L1+^qμ∗L2)≥λG˜p,q(K,L1)n+qn−qV(L1)+μG˜p,q(K,L2)n+qn−qV(L2).\frac{{\widetilde{G}}_{p,q}{\left(K,\lambda \ast {L}_{1}{\widehat{+}}_{q}\mu \ast {L}_{2})}^{\tfrac{n+q}{n-q}}}{V\left(\lambda \ast {L}_{1}{\widehat{+}}_{q}\mu \ast {L}_{2})}\ge \lambda \frac{{\widetilde{G}}_{p,q}{\left(K,{L}_{1})}^{\tfrac{n+q}{n-q}}}{V\left({L}_{1})}+\mu \frac{{\widetilde{G}}_{p,q}{\left(K,{L}_{2})}^{\tfrac{n+q}{n-q}}}{V\left({L}_{2})}.This gives inequality (1.10).According to the equality condition of inequality (3.5), we see that the equality holds in (1.10) if and only if L1{L}_{1}and L2{L}_{2}are dilates.□
Open Mathematics – de Gruyter
Published: Jan 1, 2022
Keywords: ( p , q )-mixed volume; ( p , q )-mixed geominimal surface area; monotonic inequality; isoperimetric inequality; Brunn-Minkowski-type inequality; 52A20; 52A39; 52A40
You can share this free article with as many people as you like with the url below! We hope you enjoy this feature!
Read and print from thousands of top scholarly journals.
Already have an account? Log in
Bookmark this article. You can see your Bookmarks on your DeepDyve Library.
To save an article, log in first, or sign up for a DeepDyve account if you don’t already have one.
Copy and paste the desired citation format or use the link below to download a file formatted for EndNote
Access the full text.
Sign up today, get DeepDyve free for 14 days.
All DeepDyve websites use cookies to improve your online experience. They were placed on your computer when you launched this website. You can change your cookie settings through your browser.