. . . . . .
.
. . . .
.
.
Renormalization of r
•-potentials and its
applications
Jun O’Hara (Tokyo Metropolitan University)
06/2011
Centro di Ricerca Matematica, Ennio De Giorgi, Pisa
. . . . . .
.. Motivations
Problem(PISA) : Let Ω be a triangular park. Where should we put a street light?
PISA’answer : the circumcenter (the point that makes the darkest place brightest if Ω is not obtuse).
Shibata(Fukuoka)’s answer : a point that maximizes the
“total brightness” .Definition
. .
... .
.
.
A point is an illuminating center of Ω if it maximizes Ω∋ x 7→◦
∫
Ω
dy
|x − y|2
integrated after the divergence of the integrand is cut-off
. . . . . .
.. Motivations
Problem(PISA) : Let Ω be a triangular park. Where should we put a street light?
PISA’answer : the circumcenter (the point that makes the darkest place brightest if Ω is not obtuse).
Shibata(Fukuoka)’s answer : a point that maximizes the
“total brightness” .Definition
. .
... .
.
.
A point is an illuminating center of Ω if it maximizes Ω∋ x 7→◦
∫
Ω
dy
|x − y|2
integrated after the divergence of the integrand is cut-off
. . . . . .
.. Motivations
Problem(PISA) : Let Ω be a triangular park. Where should we put a street light?
PISA’answer : the circumcenter (the point that makes the darkest place brightest if Ω is not obtuse).
Shibata(Fukuoka)’s answer : a point that maximizes the
“total brightness” .Definition
. .
... .
.
.
A point is an illuminating center of Ω if it maximizes Ω∋ x 7→◦
∫
Ω
dy
|x − y|2
integrated after the divergence of the integrand is cut-off
. . . . . .
.. Motivations
Problem(PISA) : Let Ω be a triangular park. Where should we put a street light?
PISA’answer : the circumcenter (the point that makes the darkest place brightest if Ω is not obtuse).
Shibata(Fukuoka)’s answer : a point that maximizes the
“total brightness” .Definition
. .
... .
.
.
A point is an illuminating center of Ω if it maximizes Ω∋ x 7→◦
∫
Ω
dy
|x − y|2
integrated after the divergence of the integrand is cut-off
. . . . . .
.. Outline of the first half
The renormalization used to make a family of energy functionals of knots (O ’92)
⇝ 1-parameter family of renormalized potentials
⇝ 1-parameter family of points that maximizes (or minimizes) the potentials (“generalized centers”)
One limit→ a min-max point (e.g. circumcenter of a non-obtuse triangle)
This family∋the center of mass This family∋the illuminating center
The other limit → a max-min point (e.g. incenter of a triangle) Generalization of “dual volume” in convex geometry
Uniqueness of the “centers”
. . . . . .
.. Outline of the first half
The renormalization used to make a family of energy functionals of knots (O ’92)
⇝ 1-parameter family of renormalized potentials
⇝ 1-parameter family of points that maximizes (or minimizes) the potentials (“generalized centers”)
One limit→ a min-max point (e.g. circumcenter of a non-obtuse triangle)
This family∋the center of mass This family∋the illuminating center
The other limit → a max-min point (e.g. incenter of a triangle) Generalization of “dual volume” in convex geometry
Uniqueness of the “centers”
. . . . . .
.. Outline of the first half
The renormalization used to make a family of energy functionals of knots (O ’92)
⇝ 1-parameter family of renormalized potentials
⇝ 1-parameter family of points that maximizes (or minimizes) the potentials (“generalized centers”)
One limit→ a min-max point (e.g. circumcenter of a non-obtuse triangle)
This family∋the center of mass This family∋the illuminating center
The other limit → a max-min point (e.g. incenter of a triangle) Generalization of “dual volume” in convex geometry
Uniqueness of the “centers”
. . . . . .
.. Outline of the first half
The renormalization used to make a family of energy functionals of knots (O ’92)
⇝ 1-parameter family of renormalized potentials
⇝ 1-parameter family of points that maximizes (or minimizes) the potentials (“generalized centers”)
One limit→ a min-max point (e.g. circumcenter of a non-obtuse triangle)
This family∋the center of mass This family∋the illuminating center
The other limit → a max-min point (e.g. incenter of a triangle) Generalization of “dual volume” in convex geometry
Uniqueness of the “centers”
. . . . . .
.. Outline of the first half
The renormalization used to make a family of energy functionals of knots (O ’92)
⇝ 1-parameter family of renormalized potentials
⇝ 1-parameter family of points that maximizes (or minimizes) the potentials (“generalized centers”)
One limit→ a min-max point (e.g. circumcenter of a non-obtuse triangle)
This family∋the center of mass This family∋the illuminating center
The other limit → a max-min point (e.g. incenter of a triangle) Generalization of “dual volume” in convex geometry
Uniqueness of the “centers”
. . . . . .
.. Outline of the first half
The renormalization used to make a family of energy functionals of knots (O ’92)
⇝ 1-parameter family of renormalized potentials
⇝ 1-parameter family of points that maximizes (or minimizes) the potentials (“generalized centers”)
One limit→ a min-max point (e.g. circumcenter of a non-obtuse triangle)
This family∋the center of mass This family∋the illuminating center
The other limit → a max-min point (e.g. incenter of a triangle) Generalization of “dual volume” in convex geometry
Uniqueness of the “centers”
. . . . . .
.. Outline of the first half
The renormalization used to make a family of energy functionals of knots (O ’92)
⇝ 1-parameter family of renormalized potentials
⇝ 1-parameter family of points that maximizes (or minimizes) the potentials (“generalized centers”)
One limit→ a min-max point (e.g. circumcenter of a non-obtuse triangle)
This family∋the center of mass This family∋the illuminating center
The other limit → a max-min point (e.g. incenter of a triangle) Generalization of “dual volume” in convex geometry
Uniqueness of the “centers”
. . . . . .
.. Outline of the first half
The renormalization used to make a family of energy functionals of knots (O ’92)
⇝ 1-parameter family of renormalized potentials
⇝ 1-parameter family of points that maximizes (or minimizes) the potentials (“generalized centers”)
One limit→ a min-max point (e.g. circumcenter of a non-obtuse triangle)
This family∋the center of mass This family∋the illuminating center
The other limit → a max-min point (e.g. incenter of a triangle) Generalization of “dual volume” in convex geometry
Uniqueness of the “centers”
. . . . . .
.. Outline of the first half
The renormalization used to make a family of energy functionals of knots (O ’92)
⇝ 1-parameter family of renormalized potentials
⇝ 1-parameter family of points that maximizes (or minimizes) the potentials (“generalized centers”)
One limit→ a min-max point (e.g. circumcenter of a non-obtuse triangle)
This family∋the center of mass This family∋the illuminating center
The other limit → a max-min point (e.g. incenter of a triangle) Generalization of “dual volume” in convex geometry
Uniqueness of the “centers”
. . . . . .
.. Definitions
Rm ⊃ Ω (m ≥ 2) : compact, ∂Ω : piecewise C1 .Definition
. .
... .
.
.
If α > 0 or w ̸∈ Ω
VΩ(α)(w) :=
∫
Ω
|w − z|α−mdz It is called the Riesz potential when 0 < α < m If α ≤ 0 and w ∈ Ω
VΩ(α)(w) := lim
ε→0
(∫
Ω\Bε(w)
|w − z|α−mdz+A(S
m−1
)
α ε
α
) , where A(Sm−1) is the volume of the unit sphere Sm−1 Put α1rα= log r (r > 0) if α = 0 in the formulae (although log r = lima→01a(ra− 1) in fact) When m = 2 and α = −2, Auckly-Sadun.
. . . . . .
.. Definitions
Rm ⊃ Ω (m ≥ 2) : compact, ∂Ω : piecewise C1 .Definition
. .
... .
.
.
If α > 0 or w ̸∈ Ω
VΩ(α)(w) :=
∫
Ω
|w − z|α−mdz It is called the Riesz potential when 0 < α < m If α ≤ 0 and w ∈ Ω
VΩ(α)(w) := lim
ε→0
(∫
Ω\Bε(w)
|w − z|α−mdz+A(S
m−1
)
α ε
α
) , where A(Sm−1) is the volume of the unit sphere Sm−1 Put α1rα= log r (r > 0) if α = 0 in the formulae (although log r = lima→01a(ra− 1) in fact) When m = 2 and α = −2, Auckly-Sadun.
. . . . . .
.. Definitions
Rm ⊃ Ω (m ≥ 2) : compact, ∂Ω : piecewise C1 .Definition
. .
... .
.
.
If α > 0 or w ̸∈ Ω
VΩ(α)(w) :=
∫
Ω
|w − z|α−mdz It is called the Riesz potential when 0 < α < m If α ≤ 0 and w ∈ Ω
VΩ(α)(w) := lim
ε→0
(∫
Ω\Bε(w)
|w − z|α−mdz+A(S
m−1
)
α ε
α
) , where A(Sm−1) is the volume of the unit sphere Sm−1 Put α1rα= log r (r > 0) if α = 0 in the formulae (although log r = lima→01a(ra− 1) in fact) When m = 2 and α = −2, Auckly-Sadun.
. . . . . .
.. Definitions
Rm ⊃ Ω (m ≥ 2) : compact, ∂Ω : piecewise C1 .Definition
. .
... .
.
.
If α > 0 or w ̸∈ Ω
VΩ(α)(w) :=
∫
Ω
|w − z|α−mdz It is called the Riesz potential when 0 < α < m If α ≤ 0 and w ∈ Ω
VΩ(α)(w) := lim
ε→0
(∫
Ω\Bε(w)
|w − z|α−mdz+A(S
m−1
)
α ε
α
) , where A(Sm−1) is the volume of the unit sphere Sm−1 Put α1rα= log r (r > 0) if α = 0 in the formulae (although log r = lima→01a(ra− 1) in fact) When m = 2 and α = −2, Auckly-Sadun.
. . . . . .
.. Definitions
Rm ⊃ Ω (m ≥ 2) : compact, ∂Ω : piecewise C1 .Definition
. .
... .
.
.
If α > 0 or w ̸∈ Ω
VΩ(α)(w) :=
∫
Ω
|w − z|α−mdz It is called the Riesz potential when 0 < α < m If α ≤ 0 and w ∈ Ω
VΩ(α)(w) := lim
ε→0
(∫
Ω\Bε(w)
|w − z|α−mdz+A(S
m−1
)
α ε
α
) , where A(Sm−1) is the volume of the unit sphere Sm−1 Put α1rα= log r (r > 0) if α = 0 in the formulae (although log r = lima→01a(ra− 1) in fact) When m = 2 and α = −2, Auckly-Sadun.
. . . . . .
.. Definitions
Rm ⊃ Ω (m ≥ 2) : compact, ∂Ω : piecewise C1 .Definition
. .
... .
.
.
If α > 0 or w ̸∈ Ω
VΩ(α)(w) :=
∫
Ω
|w − z|α−mdz It is called the Riesz potential when 0 < α < m If α ≤ 0 and w ∈ Ω
VΩ(α)(w) := lim
ε→0
(∫
Ω\Bε(w)
|w − z|α−mdz+A(S
m−1
)
α ε
α
) , where A(Sm−1) is the volume of the unit sphere Sm−1 Put α1rα= log r (r > 0) if α = 0 in the formulae (although log r = lima→01a(ra− 1) in fact) When m = 2 and α = −2, Auckly-Sadun.
. . . . . .
.. Properties
.Proposition .
.
... .
.
.
α <0, w ∈Ω ⇒ V◦ Ω(α)(w) = −
∫
Ωc
|w − z|α−mdz <0
Proof:
∫
Rm\Bε(w)
|w − z|α−mdz= −A(S
m−1)
α · ε
α
VΩ(α)(w)
= lim
ε→0
(∫
Ω\Bε(w)
|w − z|α−mdz−
∫
Rm\Bε(w)
|w − z|α−mdz )
= −
∫
Ωc
|w − z|α−mdz .
. . . . . .
.. Properties
.Proposition .
.
... .
.
.
α <0, w ∈Ω ⇒ V◦ Ω(α)(w) = −
∫
Ωc
|w − z|α−mdz <0
Proof:
∫
Rm\Bε(w)
|w − z|α−mdz= −A(S
m−1)
α · ε
α
VΩ(α)(w)
= lim
ε→0
(∫
Ω\Bε(w)
|w − z|α−mdz−
∫
Rm\Bε(w)
|w − z|α−mdz )
= −
∫
Ωc
|w − z|α−mdz .
. . . . . .
.. Properties
.Proposition .
.
... .
.
.
α <0, w ∈Ω ⇒ V◦ Ω(α)(w) = −
∫
Ωc
|w − z|α−mdz <0
Proof:
∫
Rm\Bε(w)
|w − z|α−mdz= −A(S
m−1)
α · ε
α
VΩ(α)(w)
= lim
ε→0
(∫
Ω\Bε(w)
|w − z|α−mdz−
∫
Rm\Bε(w)
|w − z|α−mdz )
= −
∫
Ωc
|w − z|α−mdz .
. . . . . .
.. Properties
.Proposition .
.
... .
.
.
α <0, w ∈Ω ⇒ V◦ Ω(α)(w) = −
∫
Ωc
|w − z|α−mdz <0 .Proposition
. .
... .
.
.
Ω: conex, w∈Ω ⇒ V◦ Ω(α)(w) = 1 α
∫
Sm−1
(ρ(v))α dσ(v), where ρ : Sm−1 ∋ v 7→ sup{c > 0 | w + cv ∈ Ω} ∈ R+.
ρ is called the radial function.
∫
Sm−1(ρ(v))
α dσ(v) is called the dual volume of order α (Lutwak 75,88).
A similar formula holds for general regions (after we divide Ω into convex parts)
. . . . . .
.. Properties
.Proposition .
.
... .
.
.
α <0, w ∈Ω ⇒ V◦ Ω(α)(w) = −
∫
Ωc
|w − z|α−mdz <0 .Proposition
. .
... .
.
.
Ω: conex, w∈Ω ⇒ V◦ Ω(α)(w) = 1 α
∫
Sm−1
(ρ(v))α dσ(v), where ρ : Sm−1 ∋ v 7→ sup{c > 0 | w + cv ∈ Ω} ∈ R+.
ρ is called the radial function.
∫
Sm−1(ρ(v))
α dσ(v) is called the dual volume of order α (Lutwak 75,88).
A similar formula holds for general regions (after we divide Ω into convex parts)
. . . . . .
.. Properties
.Proposition .
.
... .
.
.
α <0, w ∈Ω ⇒ V◦ Ω(α)(w) = −
∫
Ωc
|w − z|α−mdz <0 .Proposition
. .
... .
.
.
Ω: conex, w∈Ω ⇒ V◦ Ω(α)(w) = 1 α
∫
Sm−1
(ρ(v))α dσ(v), where ρ : Sm−1 ∋ v 7→ sup{c > 0 | w + cv ∈ Ω} ∈ R+.
ρ is called the radial function.
∫
Sm−1(ρ(v))
α dσ(v) is called the dual volume of order α (Lutwak 75,88).
A similar formula holds for general regions (after we divide Ω into convex parts)
. . . . . .
.. Properties
.Proposition .
.
... .
.
.
α <0, w ∈Ω ⇒ V◦ Ω(α)(w) = −
∫
Ωc
|w − z|α−mdz <0 .Proposition
. .
... .
.
.
Ω: conex, w∈Ω ⇒ V◦ Ω(α)(w) = 1 α
∫
Sm−1
(ρ(v))α dσ(v), where ρ : Sm−1 ∋ v 7→ sup{c > 0 | w + cv ∈ Ω} ∈ R+.
ρ is called the radial function.
∫
Sm−1(ρ(v))
α dσ(v) is called the dual volume of order α (Lutwak 75,88).
A similar formula holds for general regions (after we divide Ω into convex parts)
. . . . . .
.. Properties
.Proposition .
.
... .
.
.
α <0, w ∈Ω ⇒ V◦ Ω(α)(w) = −
∫
Ωc
|w − z|α−mdz <0 .Proposition
. .
... .
.
.
α≤ 0 ⇒
{ ◦
Ω∋ w → ∂Ω ⇒ VΩ(α)(w) → −∞ Ωc ∋ w → ∂Ω ⇒ VΩ(α)(w) → +∞
. . . . . .
.. Properties
.Proposition .
.
... .
.
.
α <0, w ∈Ω ⇒ V◦ Ω(α)(w) = −
∫
Ωc
|w − z|α−mdz <0 .Proposition
. .
... .
.
.
α≤ 0 ⇒
{ ◦
Ω∋ w → ∂Ω ⇒ VΩ(α)(w) → −∞ Ωc ∋ w → ∂Ω ⇒ VΩ(α)(w) → +∞
. . . . . .
.. r
•-centers
.Definition .
.
... .
.
.
An rα−m-centerof Ω is a point that attains the extreme value of VΩ(α), to be precise, a point that
minimizes VΩ(α) if α > m, maximizes VΩ(α) if 0 < α < m, maximizes
∫
Ω
− log |w − z|dz if α = m, maximizes VΩ(α)◦
Ω if α ≤ 0.
Remark that VΩ(m)(w) ≡ Vol(Ω) when α = m
. . . . . .
.. r
•-centers
.Definition .
.
... .
.
.
An rα−m-centerof Ω is a point that attains the extreme value of VΩ(α), to be precise, a point that
minimizes VΩ(α) if α > m, maximizes VΩ(α) if 0 < α < m, maximizes
∫
Ω
− log |w − z|dz if α = m, maximizes VΩ(α)◦
Ω if α ≤ 0.
Remark that VΩ(m)(w) ≡ Vol(Ω) when α = m
. . . . . .
.. r
•-centers
.Definition .
.
... .
.
.
An rα−m-centerof Ω is a point that attains the extreme value of VΩ(α), to be precise, a point that
minimizes VΩ(α) if α > m, maximizes VΩ(α) if 0 < α < m, maximizes
∫
Ω
− log |w − z|dz if α = m, maximizes VΩ(α)◦
Ω if α ≤ 0.
Remark that VΩ(m)(w) ≡ Vol(Ω) when α = m
. . . . . .
.. r
•-centers
.Definition .
.
... .
.
.
An rα−m-centerof Ω is a point that attains the extreme value of VΩ(α), to be precise, a point that
minimizes VΩ(α) if α > m, maximizes VΩ(α) if 0 < α < m, maximizes
∫
Ω
− log |w − z|dz if α = m, maximizes VΩ(α)◦
Ω if α ≤ 0.
Remark that VΩ(m)(w) ≡ Vol(Ω) when α = m
. . . . . .
.. r
•-centers
.Definition .
.
... .
.
.
An rα−m-centerof Ω is a point that attains the extreme value of VΩ(α), to be precise, a point that
minimizes VΩ(α) if α > m, maximizes VΩ(α) if 0 < α < m, maximizes
∫
Ω
− log |w − z|dz if α = m, maximizes VΩ(α)◦
Ω if α ≤ 0.
Remark that VΩ(m)(w) ≡ Vol(Ω) when α = m
. . . . . .
.. Examples
The center of mass wG of Ω is given by wG =
∫
Ωzdz
∫
Ω1dz
.
⇔
∫
Ω
(wG− z) dz = 0.
⇔ wG is a unique critical point of
∫
Ω
|w − z|2dz . It is a unique r2-center (α = m + 2)
The illuminating center of a planar domain (Shibata) is an r−2-center (m = 2, α = 0).
When m = 3, an r−1-center (α = 2) is
a “gravitational center” or a “point with highest voltage” (Herbur-Moszy´nska-Peradz´ynski)
. . . . . .
.. Examples
The center of mass wG of Ω is given by wG =
∫
Ωzdz
∫
Ω1dz
.
⇔
∫
Ω
(wG− z) dz = 0.
⇔ wG is a unique critical point of
∫
Ω
|w − z|2dz . It is a unique r2-center (α = m + 2)
The illuminating center of a planar domain (Shibata) is an r−2-center (m = 2, α = 0).
When m = 3, an r−1-center (α = 2) is
a “gravitational center” or a “point with highest voltage” (Herbur-Moszy´nska-Peradz´ynski)
. . . . . .
.. Examples
The center of mass wG of Ω is given by wG =
∫
Ωzdz
∫
Ω1dz
.
⇔
∫
Ω
(wG− z) dz = 0.
⇔ wG is a unique critical point of
∫
Ω
|w − z|2dz . It is a unique r2-center (α = m + 2)
The illuminating center of a planar domain (Shibata) is an r−2-center (m = 2, α = 0).
When m = 3, an r−1-center (α = 2) is
a “gravitational center” or a “point with highest voltage” (Herbur-Moszy´nska-Peradz´ynski)
. . . . . .
.. Examples
The center of mass wG of Ω is given by wG =
∫
Ωzdz
∫
Ω1dz
.
⇔
∫
Ω
(wG− z) dz = 0.
⇔ wG is a unique critical point of
∫
Ω
|w − z|2dz . It is a unique r2-center (α = m + 2)
The illuminating center of a planar domain (Shibata) is an r−2-center (m = 2, α = 0).
When m = 3, an r−1-center (α = 2) is
a “gravitational center” or a “point with highest voltage” (Herbur-Moszy´nska-Peradz´ynski)
. . . . . .
.. Examples
The center of mass wG of Ω is given by wG =
∫
Ωzdz
∫
Ω1dz
.
⇔
∫
Ω
(wG− z) dz = 0.
⇔ wG is a unique critical point of
∫
Ω
|w − z|2dz . It is a unique r2-center (α = m + 2)
The illuminating center of a planar domain (Shibata) is an r−2-center (m = 2, α = 0).
When m = 3, an r−1-center (α = 2) is
a “gravitational center” or a “point with highest voltage” (Herbur-Moszy´nska-Peradz´ynski)
. . . . . .
.. Examples
The center of mass wG of Ω is given by wG =
∫
Ωzdz
∫
Ω1dz
.
⇔
∫
Ω
(wG− z) dz = 0.
⇔ wG is a unique critical point of
∫
Ω
|w − z|2dz . It is a unique r2-center (α = m + 2)
The illuminating center of a planar domain (Shibata) is an r−2-center (m = 2, α = 0).
When m = 3, an r−1-center (α = 2) is
a “gravitational center” or a “point with highest voltage” (Herbur-Moszy´nska-Peradz´ynski)
. . . . . .
.. Existence and uniqueness of r
α−m-centers
.Theorem .
.
... .
. .Any compact region has an rα−m-center for any α.
ra-center is not necessarily unique. .Example
. .
... .
.
.
Ω = [−R, −1] ∪ [1, R] ⊂ R.
An ra-center (a > 1) is the origin. {r1-center} = [−1, 1].
∃ two ra-centers (a < 1, a ̸= 0) in (−R, −1) and (1, R).
. . . . . .
.. Existence and uniqueness of r
α−m-centers
.Theorem .
.
... .
. .Any compact region has an rα−m-center for any α.
ra-center is not necessarily unique. .Example
. .
... .
.
.
Ω = [−R, −1] ∪ [1, R] ⊂ R.
An ra-center (a > 1) is the origin. {r1-center} = [−1, 1].
∃ two ra-centers (a < 1, a ̸= 0) in (−R, −1) and (1, R).
. . . . . .
.. Existence and uniqueness of r
α−m-centers
.Theorem .
.
... .
. .Any compact region has an rα−m-center for any α.
ra-center is not necessarily unique. .Example
. .
... .
.
.
Ω = [−R, −1] ∪ [1, R] ⊂ R.
An ra-center (a > 1) is the origin. {r1-center} = [−1, 1].
∃ two ra-centers (a < 1, a ̸= 0) in (−R, −1) and (1, R).
. . . . . .
.. Existence and uniqueness of r
α−m-centers
.Theorem .
.
... .
. .Any compact region has an rα−m-center for any α.
ra-center is not necessarily unique. .Example
. .
... .
.
.
Ω = [−R, −1] ∪ [1, R] ⊂ R.
An ra-center (a > 1) is the origin. {r1-center} = [−1, 1].
∃ two ra-centers (a < 1, a ̸= 0) in (−R, −1) and (1, R).
. . . . . .
.. Existence and uniqueness of r
α−m-centers
.Theorem .
.
... .
. .Any compact region has an rα−m-center for any α.
ra-center is not necessarily unique. .Example
. .
... .
.
.
Ω = [−R, −1] ∪ [1, R] ⊂ R.
An ra-center (a > 1) is the origin. {r1-center} = [−1, 1].
∃ two ra-centers (a < 1, a ̸= 0) in (−R, −1) and (1, R).
. . . . . .
.. Existence and uniqueness of r
α−m-centers
.Theorem .
.
... .
. .Any compact region has an rα−m-center for any α.
ra-center is not necessarily unique. .Example
. .
... .
.
.
Ω = [−R, −1] ∪ [1, R] ⊂ R.
An ra-center (a > 1) is the origin. {r1-center} = [−1, 1].
∃ two ra-centers (a < 1, a ̸= 0) in (−R, −1) and (1, R).
. . . . . .
.. Existence and uniqueness of r
α−m-centers
.Theorem .
.
... .
. .Any compact region has an rα−m-center for any α.
ra-center is not necessarily unique. .Example
. .
... .
.
.
Ω = [−R, −1] ∪ [1, R] ⊂ R.
An ra-center (a > 1) is the origin. {r1-center} = [−1, 1].
∃ two ra-centers (a < 1, a ̸= 0) in (−R, −1) and (1, R).
. . . . . .
.. Uniqueness of r
α−m-centers
.Theorem .
.
... .
.
.
A compact region Ω has a unique rα−m-center if either α≥ m + 1,
α≤ 1 and Ω is convex. .Proof.
. .
... .
.
.
∂2VΩ(α)
∂xj2 >0 on R
m if α ≥ m + 1
∂2VΩ(α)
∂xj2 <0 on
Ω if α ≤ 1 and Ω is convex◦
.Conjecture .
.
... .
. .A convex region has a uniqe rα−m-center for any α.
. . . . . .
.. Uniqueness of r
α−m-centers
.Theorem .
.
... .
.
.
A compact region Ω has a unique rα−m-center if either α≥ m + 1,
α≤ 1 and Ω is convex. .Proof.
. .
... .
.
.
∂2VΩ(α)
∂xj2 >0 on R
m if α ≥ m + 1
∂2VΩ(α)
∂xj2 <0 on
Ω if α ≤ 1 and Ω is convex◦
.Conjecture .
.
... .
. .A convex region has a uniqe rα−m-center for any α.
. . . . . .
.. Uniqueness of r
α−m-centers
.Theorem .
.
... .
.
.
A compact region Ω has a unique rα−m-center if either α≥ m + 1,
α≤ 1 and Ω is convex. .Proof.
. .
... .
.
.
∂2VΩ(α)
∂xj2 >0 on R
m if α ≥ m + 1
∂2VΩ(α)
∂xj2 <0 on
Ω if α ≤ 1 and Ω is convex◦
.Conjecture .
.
... .
. .A convex region has a uniqe rα−m-center for any α.
. . . . . .
.. Uniqueness of r
α−m-centers
.Theorem .
.
... .
.
.
A compact region Ω has a unique rα−m-center if either α≥ m + 1,
α≤ 1 and Ω is convex. .Proof.
. .
... .
.
.
∂2VΩ(α)
∂xj2 >0 on R
m if α ≥ m + 1
∂2VΩ(α)
∂xj2 <0 on
Ω if α ≤ 1 and Ω is convex◦
.Conjecture .
.
... .
. .A convex region has a uniqe rα−m-center for any α.
. . . . . .
.. Uniqueness of r
α−m-centers
.Theorem .
.
... .
.
.
A compact region Ω has a unique rα−m-center if either α≥ m + 1,
α≤ 1 and Ω is convex. .Proof.
. .
... .
.
.
∂2VΩ(α)
∂xj2 >0 on R
m if α ≥ m + 1
∂2VΩ(α)
∂xj2 <0 on
Ω if α ≤ 1 and Ω is convex◦
.Conjecture .
.
... .
. .A convex region has a uniqe rα−m-center for any α.
. . . . . .
.. Uniqueness of r
α−m-centers
.Theorem .
.
... .
.
.
A compact region Ω has a unique rα−m-center if either α≥ m + 1,
α≤ 1 and Ω is convex. .Proof.
. .
... .
.
.
∂2VΩ(α)
∂xj2 >0 on R
m if α ≥ m + 1
∂2VΩ(α)
∂xj2 <0 on
Ω if α ≤ 1 and Ω is convex◦
.Conjecture .
.
... .
. .A convex region has a uniqe rα−m-center for any α.
. . . . . .
.. Uniqueness of r
α−m-centers
.Theorem .
.
... .
.
.
A compact region Ω has a unique rα−m-center if either α≥ m + 1,
α≤ 1 and Ω is convex. .Proof.
. .
... .
.
.
∂2VΩ(α)
∂xj2 >0 on R
m if α ≥ m + 1
∂2VΩ(α)
∂xj2 <0 on
Ω if α ≤ 1 and Ω is convex◦
.Conjecture .
.
... .
. .A convex region has a uniqe rα−m-center for any α.
. . . . . .
.. Uniqueness of r
α−m-centers
.Theorem .
.
... .
.
.
A compact region Ω has a unique rα−m-center if either α≥ m + 1,
α≤ 1 and Ω is convex. .Proof.
. .
... .
.
.
∂2VΩ(α)
∂xj2 >0 on R
m if α ≥ m + 1
∂2VΩ(α)
∂xj2 <0 on
Ω if α ≤ 1 and Ω is convex◦
.Conjecture .
.
... .
. .A convex region has a uniqe rα−m-center for any α.
. . . . . .
.. Uniqueness of r
α−m-centers
.Theorem .
.
... .
.
.
A compact region Ω has a unique rα−m-center if either α≥ m + 1,
α≤ 1 and Ω is convex. .Proof.
. .
... .
.
.
∂2VΩ(α)
∂xj2 >0 on R
m if α ≥ m + 1
∂2VΩ(α)
∂xj2 <0 on
Ω if α ≤ 1 and Ω is convex◦
.Conjecture .
.
... .
. .A convex region has a uniqe rα−m-center for any α.
. . . . . .
.. An r
α−2-center of a triangle
An rα−2-center of a triangle ∆ is a point x that gives min
x∈R2maxy∈Ω|x − y| (α → +∞). It is the circumcenter if ∆ is not obtuse
the barycenter = r2-center (α = 4)
the illuminating center = rα−2-center (α = 0) the incenter = a point x that gives max
x∈R2ymin∈Ωc|x − y|
(α → −∞)
In general, a min-max point is uniquely determined, but a max-min point is not. The latter is contained in the so-called medial axis.
. . . . . .
.. An r
α−2-center of a triangle
An rα−2-center of a triangle ∆ is a point x that gives min
x∈R2maxy∈Ω|x − y| (α → +∞). It is the circumcenter if ∆ is not obtuse
the barycenter = r2-center (α = 4)
the illuminating center = rα−2-center (α = 0) the incenter = a point x that gives max
x∈R2ymin∈Ωc|x − y|
(α → −∞)
In general, a min-max point is uniquely determined, but a max-min point is not. The latter is contained in the so-called medial axis.
. . . . . .
.. An r
α−2-center of a triangle
An rα−2-center of a triangle ∆ is a point x that gives min
x∈R2maxy∈Ω|x − y| (α → +∞). It is the circumcenter if ∆ is not obtuse
the barycenter = r2-center (α = 4)
the illuminating center = rα−2-center (α = 0) the incenter = a point x that gives max
x∈R2ymin∈Ωc|x − y|
(α → −∞)
In general, a min-max point is uniquely determined, but a max-min point is not. The latter is contained in the so-called medial axis.
. . . . . .
.. An r
α−2-center of a triangle
An rα−2-center of a triangle ∆ is a point x that gives min
x∈R2maxy∈Ω|x − y| (α → +∞). It is the circumcenter if ∆ is not obtuse
the barycenter = r2-center (α = 4)
the illuminating center = rα−2-center (α = 0) the incenter = a point x that gives max
x∈R2ymin∈Ωc|x − y|
(α → −∞)
In general, a min-max point is uniquely determined, but a max-min point is not. The latter is contained in the so-called medial axis.
. . . . . .
.. An r
α−2-center of a triangle
An rα−2-center of a triangle ∆ is a point x that gives min
x∈R2maxy∈Ω|x − y| (α → +∞). It is the circumcenter if ∆ is not obtuse
the barycenter = r2-center (α = 4)
the illuminating center = rα−2-center (α = 0) the incenter = a point x that gives max
x∈R2ymin∈Ωc|x − y|
(α → −∞)
In general, a min-max point is uniquely determined, but a max-min point is not. The latter is contained in the so-called medial axis.
. . . . . .
.. An r
α−2-center of a triangle
An rα−2-center of a triangle ∆ is a point x that gives min
x∈R2maxy∈Ω|x − y| (α → +∞). It is the circumcenter if ∆ is not obtuse
the barycenter = r2-center (α = 4)
the illuminating center = rα−2-center (α = 0) the incenter = a point x that gives max
x∈R2ymin∈Ωc|x − y|
(α → −∞)
In general, a min-max point is uniquely determined, but a max-min point is not. The latter is contained in the so-called medial axis.
. . . . . .
.. Boundary integral expression and Laplacians
.Theorem .
.
... .
.
.
If w ̸∈ ∂Ω then for any α VΩ(α)(w) = 1
α
∫
∂Ω
|w − z|α−m(z − w) · n dσ(z)
We can compute the derivatives of VΩ(α) .Proposition
. .
... .
.
.
α >2 or (α ̸= 2 and w ̸∈ ∂Ω) ⇒
∆VΩ(α)(w) = (α − 2)(α − m)VΩ(α−2)(w)
. . . . . .
.. Boundary integral expression and Laplacians
.Theorem .
.
... .
.
.
If w ̸∈ ∂Ω then for any α VΩ(α)(w) = 1
α
∫
∂Ω
|w − z|α−m(z − w) · n dσ(z)
We can compute the derivatives of VΩ(α) .Proposition
. .
... .
.
.
α >2 or (α ̸= 2 and w ̸∈ ∂Ω) ⇒
∆VΩ(α)(w) = (α − 2)(α − m)VΩ(α−2)(w)
. . . . . .
.. Boundary integral expression and Laplacians
.Theorem .
.
... .
.
.
If w ̸∈ ∂Ω then for any α VΩ(α)(w) = 1
α
∫
∂Ω
|w − z|α−m(z − w) · n dσ(z)
We can compute the derivatives of VΩ(α) .Proposition
. .
... .
.
.
α >2 or (α ̸= 2 and w ̸∈ ∂Ω) ⇒
∆VΩ(α)(w) = (α − 2)(α − m)VΩ(α−2)(w)
. . . . . .
.. Boundary integral expression and Laplacians
.Theorem .
.
... .
.
.
If w ̸∈ ∂Ω then for any α VΩ(α)(w) = 1
α
∫
∂Ω
|w − z|α−m(z − w) · n dσ(z)
We can compute the derivatives of VΩ(α) .Proposition
. .
... .
.
.
α >2 or (α ̸= 2 and w ̸∈ ∂Ω) ⇒
∆VΩ(α)(w) = (α − 2)(α − m)VΩ(α−2)(w)
. . . . . .
.. Minimal unfolded region
.Definition .
.
... .
.
.
Mv := sup
x∈Ω
x· v (v ∈ Sm−1),
Ω+v,b:= Ω ∩ {x ∈ Rm| x · v > b} (b ∈ R), Reflv,b:= a reflection in {x ∈ Rm| x · v = b}.
lv := inf{a | a ≤ Mv,Reflv,b(Ω+v,b) ⊂ Ω (a ≤ ∀b ≤ Mv)}. Define the minimal unfolded region of Ω by
U f(Ω) = ∩
v∈Sm−1
{x ∈ Rm| x · v ≤ lv} .
. . . . . .
.. Examples of minimal unfolded regions
U f(Ω) ̸⊂ Ω
. . . . . .
.. Examples of minimal unfolded regions
U f(Ω) ̸⊂ Ω
. . . . . .
.. Examples of minimal unfolded regions
U f(Ω) ̸⊂ Ω
. . . . . .
.. Centers in minimal unfolded regions
The integrands of our potentials and their derivatives are quite symmetric.
The moving plane method (Gidas-Ni-Nirenberg) implies .Theorem
. .
... .
. .Any rα−m-center of Ω is in the minimal unfolded region of Ω.