3
$\begingroup$

I've got a real-valued function of several vectors $f(u,v,w)$ formed by taking scalar products of linear combinations of the vectors,

I want to Taylor expand around small $v$ by writing

$$f(u,\delta v, w) = A + B\delta + O(\delta^2)$$

for small real $\delta$. When I do this in Mathematica, the output gives me terms like $v(1.u)$ rather than simplifying this to $v.u$.

Note that I have put into my assumptions that $u,v,w$ are all vectors. My Mathematica code is below.

$Assumptions = (p|q|r|P|m|n) ∈ Vectors[4]
(p|q|r|P|m|n) ∈ Vectors[4, Complexes]
FullSimplify[
  TensorExpand[
    Series[((n + q).(n + q)(p + q + δ r).(p + q + δ r))/
     (P.P(p + δ r).(p + δ r) - 
       (p + n + δ r).(p + n + δ r)(p + q + δ r).(p + q + δ r)),
     {δ,0,1}]],
  {p.p == 0, q.q == 0, r.r == 0, n.n == 0, m.m == 0, δ ∈ Reals}]
-(n.q/n.p) + 
  (r n.q (2 (1.n + 1.p + n.1 + p.1) p.q - (1.p + p.1) P.P) δ)/(4 (n.p)^2 p.q) + O[δ]^2
$\endgroup$
1
  • 1
    $\begingroup$ To get what you want you need to put the TensorExpand inside the Series as Series[TensorExpand[<tensor expression>]]. $\endgroup$ Commented May 16, 2014 at 16:02

1 Answer 1

3
$\begingroup$

Since it solved the problem, I'll promote my comment to an "official" answer.

To get the behaviour that you require you need to put the TensorExpand inside the Series thus:

$Assumptions = (p | q | r | P | m | n) \[Element] Vectors[4]

(* (p | q | r | P | m | n) \[Element] Vectors[4, Complexes] *)

FullSimplify[
  Series[TensorExpand[((n + q).(n + q) (p + q + \[Delta] r).(p + 
     q + \[Delta] r))/(P.P (p + \[Delta] r).(p + \[Delta] r) - (p \
     + n + \[Delta] r).(p + n + \[Delta] r) (p + q + \[Delta] r).(p + 
     q + \[Delta] r))], {\[Delta], 0, 1}], {p.p == 0, q.q == 0, 
     r.r == 0, n.n == 0, m.m == 0, \[Delta] \[Element] Reals}]

(* SeriesData[\[Delta], 0, {-Dot[n, p]^(-1) Dot[n, q], 
Rational[1, 2] Dot[n, p]^(-2) Dot[n, q] Dot[p, q]^(-1) (
2 Dot[p, q] (Dot[n, r] + Dot[p, r]) - Dot[p, r] Dot[P, P])}, 0, 2, 1] *)
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.