B Interpreting Path Ordered Exponentials with Non-Sensical Integration Variables

  • B
  • Thread starter Thread starter etotheipi
  • Start date Start date
  • Tags Tags
    Path
etotheipi
I see a term like$$\int_0^{t''} \left[ \int_{t'}^t \omega(t'') \omega(t') dt' \right] dt''$$here ##\omega## is a matrix. How to interpret this integral - the integration variables are in the limits, in places where they don't make sense. Is that a mistake? It's given that the range of integration is ##0 < t' < t'' < t##, i.e. a triangle in the ##t'##-##t''## plane. Reference is top of page 50 here. Thanks!
 
Physics news on Phys.org
I think ##\omega\, : \,[0,1] \longrightarrow \operatorname{SO}(3)## is simply a path on the manifold. The point is to demonstrate why general, i.e. non-commuting matrices cannot be handled the same way as (commuting) scalars. The "let's pretend as if" part is a bit confusing, since he uses the same variables.
 
  • Like
Likes etotheipi
Hope it will be of some use

formula (3.11) at page 49 means that ##\omega(t)\in T_e(SO(3))##

formula (3.16) follows from the standard successive procedure of the contraction mapping principle for the equation ##R=E+\int_0^t\omega Rd\xi##.
 
  • Informative
Likes etotheipi
That's interesting! Maybe a bit above my level of maths. Let me see what I understand; by the contraction mapping principle, the successive approximations to ##R(t)## are $$R_n(t) = 1 + \int_0^t \omega(\xi) R_{n-1}(\xi) d\xi$$So for instance$$R_1(t) = 1 + \int_0^t \omega(\xi) d\xi$$and the next approximation is$$\begin{align*}

R_2(t) = 1 + \int_0^t \omega(\xi) R_1(\xi) d\xi &= 1+ \int_0^t \omega(\xi) \left[ 1 + \int_0^t \omega(\eta) d\eta \right] d\xi \\

&= 1 + \int_0^t \omega(\xi) d\xi + \int_0^t \int_0^t \omega(\xi) \omega(\eta) d\eta d\xi

\end{align*}$$and then continue on and take ##R(t) = \lim_{n \rightarrow \infty}R_n(t)##. Except, I have some different limits in the second integral here to Tong... 🤭
 
Last edited by a moderator:
You write ##R_1(\xi)## it is correct ; but then you must substitute ##R_1(\xi)=1+\int_0^\xi\omega(s)ds##
 
  • Like
Likes etotheipi
wrobel said:
You write ##R_1(\xi)## it is correct ; but then you must substitute ##R_1(\xi)=1+\int_0^\xi\omega(s)ds##

Ah, yes that's it, thanks ☺
 
Back
Top