Hyperquadrics

Definition

Definition
The kernel of a quadratic affine form H : A −→ K is the set

Ker (H) := H −1 (0) = {P ∈ A| H(P) = 0} .

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 1 / 14

Definition
A subset H ⊆ A is called a quadratic manifold or hyperquadric if there is
a quadratic affine form H : A −→ K such that H = Ker (H).
A quadratic manifold in an affine plane A2 is called a conic(or curve of
second degree).

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 2 / 14

Definition
A subset H ⊆ A is called a quadratic manifold or hyperquadric if there is
a quadratic affine form H : A −→ K such that H = Ker (H).
A quadratic manifold in an affine plane A2 is called a conic(or curve of
second degree).
A quadratic manifold in a 3−dimensional affine space A3 is called a
quadric(or surface of second degree).

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 2 / 14

Definition
A subset H ⊆ A is called a quadratic manifold or hyperquadric if there is
a quadratic affine form H : A −→ K such that H = Ker (H).
A quadratic manifold in an affine plane A2 is called a conic(or curve of
second degree).
A quadratic manifold in a 3−dimensional affine space A3 is called a
quadric(or surface of second degree).
For a hyperquadric H = Ker (H), the equality

H(P) = 0

is called the equation of the hyperquadric H.

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 2 / 14

Definition
A subset H ⊆ A is called a quadratic manifold or hyperquadric if there is
a quadratic affine form H : A −→ K such that H = Ker (H).
A quadratic manifold in an affine plane A2 is called a conic(or curve of
second degree).
A quadratic manifold in a 3−dimensional affine space A3 is called a
quadric(or surface of second degree).
For a hyperquadric H = Ker (H), the equality

H(P) = 0

is called the equation of the hyperquadric H.

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 2 / 14

Remark
If O ∈ A is a fixed point, and r = OP, with P ∈ H, the equation of the
hyperquadric can be written as

h(r ) + 2f (r ) + c = 0 ,

where h and f are respectively the quadratic form and the linear form
associated to the affine quadratic form H, and c = H(O).

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 3 / 14

Remark
If A is finite-dimensional, and F = (O; B = (vi )i=1,n ) is a cartesian
frame in A, then P(xi ) ∈ H if and only if
n
X n
X
aij xi xj + 2 bi xi + c = 0 ,
i,j=1 i=1

with aij = aji , (∀)i, j = 1, n. In matrix form this can be written as

X t · A · X + 2B · X + c = 0 ,

with A = At , or

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 4 / 14

Remark
If A is finite-dimensional, and F = (O; B = (vi )i=1,n ) is a cartesian
frame in A, then P(xi ) ∈ H if and only if
n
X n
X
aij xi xj + 2 bi xi + c = 0 ,
i,j=1 i=1

with aij = aji , (∀)i, j = 1, n. In matrix form this can be written as

X t · A · X + 2B · X + c = 0 ,

with A = At , or

A Bt
   
X
Xt
 
1 · · = 0.
B c 1

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 4 / 14

Remark
If A is finite-dimensional, and F = (O; B = (vi )i=1,n ) is a cartesian
frame in A, then P(xi ) ∈ H if and only if
n
X n
X
aij xi xj + 2 bi xi + c = 0 ,
i,j=1 i=1

with aij = aji , (∀)i, j = 1, n. In matrix form this can be written as

X t · A · X + 2B · X + c = 0 ,

with A = At , or

A Bt
   
X
Xt
 
1 · · = 0.
B c 1

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 4 / 14

Definition
Let H ⊆ A be a hyperquadric of equation H(P) = 0.
A point P0 ∈ A is called a center of the hyperquadric H if it is a center
of the quadratic affine form H.

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 5 / 14

Definition
Let H ⊆ A be a hyperquadric of equation H(P) = 0.
A point P0 ∈ A is called a center of the hyperquadric H if it is a center
of the quadratic affine form H.
A point P0 ∈ H is called a singular point of H if it is a center of H.

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 5 / 14

Definition
Let H ⊆ A be a hyperquadric of equation H(P) = 0.
A point P0 ∈ A is called a center of the hyperquadric H if it is a center
of the quadratic affine form H.
A point P0 ∈ H is called a singular point of H if it is a center of H. A
point P0 ∈ H is called an ordinary point of H if it is not a singular
point.

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 5 / 14

Definition
Let H ⊆ A be a hyperquadric of equation H(P) = 0.
A point P0 ∈ A is called a center of the hyperquadric H if it is a center
of the quadratic affine form H.
A point P0 ∈ H is called a singular point of H if it is a center of H. A
point P0 ∈ H is called an ordinary point of H if it is not a singular
point.

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 5 / 14

Remark
P0 ∈ A is a center of H if
n
X
aij xj + bi = 0 , (∀)i = 1, n .
j=1

P0 ∈ A is a singular point of H if
 n
P


 aij xj + bi = 0 , (∀)i = 1, n
 j=1



 n
P
bi xi + c = 0



i=1

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 6 / 14

Remark
P0 ∈ A is a center of H if
n
X
aij xj + bi = 0 , (∀)i = 1, n .
j=1

P0 ∈ A is a singular point of H if
 n
P


 aij xj + bi = 0 , (∀)i = 1, n
 j=1



 n
P
bi xi + c = 0



i=1

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 6 / 14

Remark
A hyperquadric H has centers if and only if

rank(A) = rank(A B t ) ⇐⇒ ρ = r or ρ = r + 1 .

A hyperquadric H has singular points if and only if

rank(A) = rank(A B t ) = rank(D) ⇐⇒ ρ = r .

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 7 / 14

Remark
A hyperquadric H has centers if and only if

rank(A) = rank(A B t ) ⇐⇒ ρ = r or ρ = r + 1 .

A hyperquadric H has singular points if and only if

rank(A) = rank(A B t ) = rank(D) ⇐⇒ ρ = r .

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 7 / 14

Proposition
Let An be a finite-dimensional real affine space, and H ⊆ An a
hyperquadric. Then there is a cartesian frame F = (O; B = (vi )i=1,n ) of
An such that the equation of H has one of the following forms:
a) x12 + · · · + xp2 − xp+1
2 − · · · − xr2 = 0, with p ≥ 2r , if ρ = r .
b) x1 + · · · + xp − xp+1 − · · · − xr2 = 1, if ρ = r + 1.
2 2 2

c) x12 + · · · + xp2 − xp+1

2 − · · · − xr2 = 2xr +1 , with p ≥ 2r , if ρ = r + 2.

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 8 / 14

Proposition
Let An be a finite-dimensional real affine space, and H ⊆ An a
hyperquadric. Then there is a cartesian frame F = (O; B = (vi )i=1,n ) of
An such that the equation of H has one of the following forms:
a) x12 + · · · + xp2 − xp+1
2 − · · · − xr2 = 0, with p ≥ 2r , if ρ = r .
b) x1 + · · · + xp − xp+1 − · · · − xr2 = 1, if ρ = r + 1.
2 2 2

c) x12 + · · · + xp2 − xp+1

2 − · · · − xr2 = 2xr +1 , with p ≥ 2r , if ρ = r + 2.

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 8 / 14

 Affine classification of real conics
Let A2 be a real affine plane, H ⊆ A2 a conic, and F = (O; e1 , e2 ) a
cartesian frame in A2 . Then there are real scalars
a11 , a12 , a22 , b1 , b2 , c ∈ R such that

H = {P(x, y ) ∈ A| H(P) = 0 ,

where
H(P) = a11 x 2 + 2a12 xy + a22 y 2 + 2b1 x + 2b2 y + c .

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 9 / 14

The matrices associated to H are
 
  a11 a12 b1
a11 a12
A= and D =  a12 a22 b2  .
a12 a22
b1 b2 c

We shall use the standard notations r = rank(A), ρ = rank(D),

δ = det(A), ∆ = det(D). The conic H is called
a) a non-singular conic if ∆ 6= 0.

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 10 / 14

The matrices associated to H are
 
  a11 a12 b1
a11 a12
A= and D =  a12 a22 b2  .
a12 a22
b1 b2 c

We shall use the standard notations r = rank(A), ρ = rank(D),

δ = det(A), ∆ = det(D). The conic H is called
a) a non-singular conic if ∆ 6= 0.
b) a singular conic if ∆ = 0.

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 10 / 14

The matrices associated to H are
 
  a11 a12 b1
a11 a12
A= and D =  a12 a22 b2  .
a12 a22
b1 b2 c

We shall use the standard notations r = rank(A), ρ = rank(D),

δ = det(A), ∆ = det(D). The conic H is called
a) a non-singular conic if ∆ 6= 0.
b) a singular conic if ∆ = 0.
b’) a singular degenerated conic if ∆ = 0 and δ = 0.

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 10 / 14

The matrices associated to H are
 
  a11 a12 b1
a11 a12
A= and D =  a12 a22 b2  .
a12 a22
b1 b2 c

We shall use the standard notations r = rank(A), ρ = rank(D),

δ = det(A), ∆ = det(D). The conic H is called
a) a non-singular conic if ∆ 6= 0.
b) a singular conic if ∆ = 0.
b’) a singular degenerated conic if ∆ = 0 and δ = 0.

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 10 / 14

The types of the conics that can exist are given in the following table,
classified in terms of the invariants ∆, δ, ρ, r and p:
nature of H ρ r p normal equation name of H
∆ 6= 0 3 2 2 x2 + y2 = 1 δ > 0. Ellipse
non-singular 1 x2 − y2 = 1 δ < 0. Hyperbola
conic 0 −x 2 − y 2 = 1 δ > 0. Empty ellipse
1 1 x 2 = 2y δ = 0. Parabola
∆ = 0, δ 6= 0 2 2 2 x2 + y2 = 0 δ > 0. Double point
singular conic 1 x2 − y2 = 0 δ < 0. Pair of intersecting lines
∆ = 0, δ = 0 2 1 1 x2 = 1 Pair of parallel lines
singular degenerated 0 −x 2 = 1 Empty pair of lines
conic 1 1 1 x2 = 0 Double line

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 11 / 14

 Affine classification of real quadrics
Let A3 be a 3−dimensional real affine space, H ⊆ A3 a quadric, and
F = (O; e1 , e2 , e3 ) a cartesian frame in A3 . Then there are real scalars
aij , bi , c ∈ R, i, j = 1, 3, such that

H = {P(x, y , z)| H(P) = 0} ,

where
H(P) = a11 x 2 + a22 y 2 + a33 z 2 + 2a12 xy + 2a13 xz + 2a23 yz+
+2b1 x + 2b2 y + 2b3 z + c .

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 12 / 14

The matrices associated to H are
 
  a11 a12 a13 b1
a11 a12 a13  a12 a22 a23 b2 
A =  a12 a22 a23  and D = 
 a13 a23 a33 b3  .

a13 a23 a33
b1 b2 b3 c

Using the standard notations r = rank(A), ρ = rank(D), δ = det(A),

∆ = det(D), the quadric H is called
a) a non-singular quadric if ∆ 6= 0.

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 13 / 14

The matrices associated to H are
 
  a11 a12 a13 b1
a11 a12 a13  a12 a22 a23 b2 
A =  a12 a22 a23  and D = 
 a13 a23 a33 b3  .

a13 a23 a33
b1 b2 b3 c

Using the standard notations r = rank(A), ρ = rank(D), δ = det(A),

∆ = det(D), the quadric H is called
a) a non-singular quadric if ∆ 6= 0.
b) a singular quadric if ∆ = 0.

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 13 / 14

The matrices associated to H are
 
  a11 a12 a13 b1
a11 a12 a13  a12 a22 a23 b2 
A =  a12 a22 a23  and D = 
 a13 a23 a33 b3  .

a13 a23 a33
b1 b2 b3 c

Using the standard notations r = rank(A), ρ = rank(D), δ = det(A),

∆ = det(D), the quadric H is called
a) a non-singular quadric if ∆ 6= 0.
b) a singular quadric if ∆ = 0.
b’) a singular degenerated quadric if ∆ = 0 and ρ ≤ 2.

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 13 / 14

The matrices associated to H are
 
  a11 a12 a13 b1
a11 a12 a13  a12 a22 a23 b2 
A =  a12 a22 a23  and D = 
 a13 a23 a33 b3  .

a13 a23 a33
b1 b2 b3 c

Using the standard notations r = rank(A), ρ = rank(D), δ = det(A),

∆ = det(D), the quadric H is called
a) a non-singular quadric if ∆ 6= 0.
b) a singular quadric if ∆ = 0.
b’) a singular degenerated quadric if ∆ = 0 and ρ ≤ 2.
The types of the quadrics that can exist are given in the following table,
classified in terms of the invariants ∆, δ, ρ, r and p:

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 13 / 14

The matrices associated to H are
 
  a11 a12 a13 b1
a11 a12 a13  a12 a22 a23 b2 
A =  a12 a22 a23  and D = 
 a13 a23 a33 b3  .

a13 a23 a33
b1 b2 b3 c

Using the standard notations r = rank(A), ρ = rank(D), δ = det(A),

∆ = det(D), the quadric H is called
a) a non-singular quadric if ∆ 6= 0.
b) a singular quadric if ∆ = 0.
b’) a singular degenerated quadric if ∆ = 0 and ρ ≤ 2.
The types of the quadrics that can exist are given in the following table,
classified in terms of the invariants ∆, δ, ρ, r and p:

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 13 / 14

 nature of H ρ r p normal equation name of H
∆ 6= 0 4 3 3 x2 + y2 + z2 = 1 Ellipsoid
non-singular 6 0)
(δ = 2 x2 + y2 − z2 = 1 Hyperboloid with one surface
quadric 1 x2 − y2 − z2 = 1 Hyperboloid with two surfaces
0 −x 2 − y 2 − z 2 = 1 Empty ellipsoid
2 2 x 2 + y 2 = 2z Elliptic paraboloid
(δ 6= 0) 1 x 2 − y 2 = 2z Hyperbolic paraboloid
∆=0 3 3 3 x2 + y2 + z2 = 0 Double point
singular (δ 6= 0) 2 x2 + y2 − z2 = 0 Cone
non-degenerated 2 2 x2 + y2 = 1 Ellyptic cylinder
quadric (δ = 0) 1 x2 − y2 = 1 Hyperbolic cylinder
0 −x 2 − y 2 = 1 Empty cylinder
1(δ = 0) 1 x 2 = 2y Parabolic cylinder
∆ = 0, ρ ≤ 2 2 2 2 x2 + y2 = 0 Double line
singular 1 x2 − y2 = 0 Pair of intersecting planes
degenerated 1 1 x2 = 1 Pair of parallel planes
quadric 0 −x 2 = 1 Empty pair of planes
1 1 1 x2 = 0 Double plane

Lect.dr. M.Chiş () Lecture 8 7.IV.2009 14 / 14