Analyzing Commutative Exponential of Matrix in Antisymmetric Dyadics, Exercises of Classical and Relativistic Mechanics

The properties of the commutative exponential of a matrix x in the space of antisymmetric dyadics, denoted o(n). The author derives the basis of o(n) and shows that the matrix x is determined by its components xi,j for i < j. The document also discusses the relationship between the matrix x and the vector a, which uniquely represents an element x of o(n). Furthermore, the document explains how the vector f, obtained from the normalised dot product a · j, is related to the gradient vector ∇qg and the vector vf [g]. The document concludes by demonstrating the equality of the matrix exp(tvf )(q,p) and etxa(q,p) for all (q,p) in rn × rn.

Typology: Exercises

2011/2012

Uploaded on 07/19/2012

davdas900
davdas900 🇮🇳

4.6

(5)

110 documents

1 / 2

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
n!=
X
n=0
(tX)n
n!=etX
.
Since
tX
and
tX
commute, I get
etX (etX )=etX etX =etXtX =e0= 1
. Thus,
etX O(n)
. Of course,
e0X= 1
. Dierentiating termwise, as is allowed for entire functions on Banach algebras,
d
dt etX =
X
n=0
ntn1Xn
n!=
X
n=1
(tX)n1X
(n1)! =X
X
n=0
(tX)n
n!=XetX
,
which at
t= 0
is
X1 = X
.
3
If
Xo(n)
, then
Xi,j =Xj,i
for indices
i, j
. If
i=j
, it follows that
Xi,j = 0
. If
i > j
, then
Xi,j
is de-
termined by
Xj,i
, where
j < i
. Thus,
X
is determined by
Xi,j
for
i < j
. Conversely, given any values for
Xi,j
for
i < j
, let
Xi,j
be
Xj,i
when
i > j
and let
Xi,j
be
0
when
i=j
. Then
Xo(n)
. Thus, a basis of
o(n)
is
{Ei,j
.
.
.
i < j}
, where
(Ei,j)i,j = 1
,
(Ei,j)j,i =1
, and every other component of
Ei,j
is
0
. (That is,
(Ei,j)k ,l := δi,j
k,l
.) When
n= 3
, we have
{E1,2, E1,3, E2,3}
. Since
X1=E2,3
,
X2=E1,3
, and
X3=E1,2
,
another basis is
{X1, X2, X3}
.
4
If
v= (v1, v2, v3)
, then
(a·X)v=
0a3a2
a30a1
a2a10

v1
v2
v3
=
a2v3a3v2
a3v1a1v3
a1v2a2v1
=a×v
.
Thus, the
t
derivative of
eta·Xv
is
(a·X)eta·Xv=a×eta·Xv
.
5
Let
G
be any smooth function on the manifold
X
. Then
vF[G] = {F, G}=∂F
∂p1
∂G
∂q1
∂F
∂q1
∂G
∂p1
+∂F
∂p2
∂G
∂q2
∂F
∂q2
∂G
∂p2
+∂F
∂p3
∂G
∂q3
∂F
∂q3
∂G
∂p3
= (a2q3a3q2)∂G
∂q1
(a3p2a2p3)∂G
∂p1
+ (a3q1a1q3)∂G
∂q2
(a1p3a3p1)∂G
∂p2
+ (a1q2a2q1)∂G
∂q3
(a2p1a1p2)∂G
∂p3
= (a×q)· qG+ (a×p)· pG
.
Thus,
vF= (a×q)· q+ (a×p)· p
. Identifying the tangent spaces to
R3×R3
with
R3×R3
itself, this
becomes
vF= (a×q,a×p) = (a,a) (×,×) (q,p)
.
6
At
t= 0
, both sides of equation (1) reduce to
(q,p)
. The
t
derivative of the left side is
vF|exp (tvF)(q,p)=
(a,a) (×,×) exp (tvF)(q,p)
; the
t
derivative of the right side is
(a×eta·Xq,a×eta·Xp) = (a,a) (×,×)
eta·X(q,p)
. Thus, both sides satisfy the same initial value problem, so they are equal.
Page 1 of 2
docsity.com
pf2

Partial preview of the text

Download Analyzing Commutative Exponential of Matrix in Antisymmetric Dyadics and more Exercises Classical and Relativistic Mechanics in PDF only on Docsity!

n!

X

n=

(−tX)n n!

= e−tX^.

Since tX and −tX commute, I get etX^ (etX^ )∗^ = etX^ e−tX^ = etX−tX^ = e^0 = 1. Thus, etX^ ∈ O(n). Of course, e^0 X^ = 1. Dierentiating termwise, as is allowed for entire functions on Banach algebras,

d dt

etX^ =

X

n=

ntn−^1 Xn n!

X

n=

(tX)n−^1 X (n − 1)!

= X

X

n=

(tX)n n!

= XetX^ ,

which at t = 0 is X1 = X.

3 If X ∈ o(n), then Xi,j = −Xj,i for indices i, j. If i = j, it follows that Xi,j = 0. If i > j, then Xi,j is de- termined by Xj,i, where j < i. Thus, X is determined by Xi,j for i < j. Conversely, given any values for Xi,j for i < j, let Xi,j be −Xj,i when i > j and let Xi,j be 0 when i = j. Then X ∈ o(n). Thus, a basis of o(n) is {Ei,j^ ... i < j}, where (Ei,j )i,j = 1, (Ei,j )j,i = − 1 , and every other component of Ei,j is 0. (That is, (Ei,j )k,l := δi,jk,l.) When n = 3, we have {E 1 , 2 , E 1 , 3 , E 2 , 3 }. Since X 1 = −E 2 , 3 , X 2 = E 1 , 3 , and X 3 = −E 1 , 2 , another basis is {X 1 , X 2 , X 3 }.

4 If v = (v 1 , v 2 , v 3 ), then

(a · X)v =

0 −a 3 a 2 a 3 0 −a 1 −a 2 a 1 0

v 1 v 2 v 3

a 2 v 3 − a 3 v 2 a 3 v 1 − a 1 v 3 a 1 v 2 − a 2 v 1

= a × v.

Thus, the t derivative of eta·Xv is (a · X)eta·Xv = a × eta·Xv.

5 Let G be any smooth function on the manifold X. Then

vF [G] = {F, G} = ∂F ∂p 1

∂G

∂q 1

− ∂F

∂q 1

∂G

∂p 1

+ ∂F

∂p 2

∂G

∂q 2

− ∂F

∂q 2

∂G

∂p 2

+ ∂F

∂p 3

∂G

∂q 3

− ∂F

∂q 3

∂G

∂p 3

= (a 2 q 3 − a 3 q 2 )

∂G

∂q 1

− (a 3 p 2 − a 2 p 3 )

∂G

∂p 1

  • (a 3 q 1 − a 1 q 3 )

∂G

∂q 2

− (a 1 p 3 − a 3 p 1 ) ∂G ∂p 2

  • (a 1 q 2 − a 2 q 1 ) ∂G ∂q 3

− (a 2 p 1 − a 1 p 2 ) ∂G ∂p 3 = (a × q) · ∇qG + (a × p) · ∇pG.

Thus, vF = (a × q) · ∇q + (a × p) · ∇p. Identifying the tangent spaces to R^3 × R^3 with R^3 × R^3 itself, this becomes vF = (a × q, a × p) = (a, a) (×, ×) (q, p).

6 At t = 0, both sides of equation (1) reduce to (q, p). The t derivative of the left side is vF |exp (tvF )(q,p) = (a, a) (×, ×) exp (tvF )(q, p); the t derivative of the right side is (a × eta·Xq, a × eta·Xp) = (a, a) (×, ×) eta·X(q, p). Thus, both sides satisfy the same initial value problem, so they are equal.

Page 1 of 2

docsity.com

7 Recall that we have the basis {Ei,j^ ... i < j} for o(n). Then an element X of o(n) is given uniquely by a 2 vector a in Λ^2 Rn, under X =

P

i<j ai,j^ Ei,j^. Symbolically, I ll indicate this as^ Xa. Then^ Xav^ =^ a^ ·^ v, where a is treated as an antisymmetric dyadic for purposes of the dot product. The t derivative of etXa^ v is XaetXa^ v = a · etXa^ v. If (q, p) ∈ Rn^ × Rn, then let J be the 2 vector q ∧ p. Then let F be the normalised dot product a · J. That is, F =

P

i<j ai,j^ (qipj^ −^ qj^ pi). Given G ∈ C∞(Rn^ × Rn),

vF [G] = {F, G} =

X

k

∂F

∂pk

∂G

∂qk

− ∂F

∂qk

∂G

∂pk

X

k

X

i<j

ai,j (qiδj,k − qj δi,k) ∂G ∂qk

− ai,j (δi,kpj − δj,kpi) ∂G ∂pk

X

i<j

ai,j

qi

∂G

∂qj^ −^ qj

∂G

∂qi^ +^ pi

∂G

∂pj^ −^ pj

∂G

∂pi

X

i,j

ai,j

qi

∂G

∂qj^ +^ pi

∂G

∂pj

= (a · q) · ∇qG + (a · p) · ∇pG.

Thus, vF = (a · q) · ∇q + (a · p) · ∇p = (a · q, a · p) = (a, a) (·, ·) (q, p). In this context, equation (1) becomes

exp (tvF )(q, p) = etXa^ (q, p).

Again, at t = 0, both sides become (q, p). The t derivative of the left side is (a, a) (·, ·) exp (tvF )(q, p), and that of the right side is (a, a) (·, ·) etXa^ (q, p). Therefore, the two sides are still equal.

Page 2 of 2

docsity.com