Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.soft-sys.math.mathematica > #5644 > unrolled thread

Re: Bug in Series with NonCommutativeProduct?

Started byDaniel Lichtblau <danl@wolfram.com>
First post2011-10-14 09:59 +0000
Last post2011-10-20 11:46 +0000
Articles 2 — 2 participants

Back to article view | Back to comp.soft-sys.math.mathematica

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.


Contents

  Re: Bug in Series with NonCommutativeProduct? Daniel Lichtblau <danl@wolfram.com> - 2011-10-14 09:59 +0000
    Re: Bug in Series with NonCommutativeProduct? D J G C <dundjoh@googlemail.com> - 2011-10-20 11:46 +0000

#5644 — Re: Bug in Series with NonCommutativeProduct?

FromDaniel Lichtblau <danl@wolfram.com>
Date2011-10-14 09:59 +0000
SubjectRe: Bug in Series with NonCommutativeProduct?
Message-ID<j79160$i4b$1@smc.vnet.net>
On 10/13/2011 02:48 AM, D J G C wrote:
>> Series[(A h) ** (B h), {h, 0, 2}]
>
> ... + A B  h^2+O[h]^3
>
> The problem is the appearance of a commutative product A B. It should
> instead be A ** B
>
> However applying the D gives the correct answer:
>
>> D[(A h ) ** (B h), {h, 2}]
>
> 0**(B h)+2 A**B+(A h)**0

It's just doing a "blind" Taylor expansion. One can see this by 
replacing NonCommutativeMultiply with some (undefined) function f:

In[242]:= Series[f[(A h), (B h)], {h, 0, 2}]

Out[242]= SeriesData[h, 0, {
f[0, 0], B Derivative[0, 1][f][0, 0] + A Derivative[1, 0][f][0, 0],
   Rational[1, 2] (
    B^2 Derivative[0, 2][f][0, 0] + 2 A B Derivative[1, 1][f][
      0, 0] + A^2 Derivative[2, 0][f][0, 0])}, 0, 3, 1]


Daniel Lichtblau
Wolfram Research

[toc] | [next] | [standalone]


#5731

FromD J G C <dundjoh@googlemail.com>
Date2011-10-20 11:46 +0000
Message-ID<j7p1nf$5ju$1@smc.vnet.net>
In reply to#5644
On Oct 14, 11:59 am, Daniel Lichtblau <d...@wolfram.com> wrote:
> On 10/13/2011 02:48 AM, D J G C wrote:
>
> >>Series[(A h) ** (B h), {h, 0, 2}]
>
> > ... + A B  h^2+O[h]^3
>
> > The problem is the appearance of a commutative product A B. It should
> > instead be A ** B
>
> > However applying the D gives the correct answer:
>
> >> D[(A h ) ** (B h), {h, 2}]
>
> > 0**(B h)+2 A**B+(A h)**0
>
> It's just doing a "blind" Taylor expansion. One can see this by
> replacing NonCommutativeMultiply with some (undefined) function f:
>
> In[242]:=Series[f[(A h), (B h)], {h, 0, 2}]
>
> Out[242]= SeriesData[h, 0, {
> f[0, 0], B Derivative[0, 1][f][0, 0] + A Derivative[1, 0][f][0, 0],
>    Rational[1, 2] (
>     B^2 Derivative[0, 2][f][0, 0] + 2 A B Derivative[1, 1][f][
>       0, 0] + A^2 Derivative[2, 0][f][0, 0])}, 0, 3, 1]
>
> Daniel Lichtblau
> Wolfram Research

A series always is a "blind" Taylor expansion.

The bug is, that in this "blind" expansion it gets the second
derivative wrong, by ignoring the possibility for a non-commutative
product.

The D[] is aware of this. Why not use its result in the expansion
(which by the way is the definition of a series)?

This is a bug, as Mathematica is a mathematical software, offering the
(a) non-commutative product, which part of it does not handle. At
least the user has to be warned.

[toc] | [prev] | [standalone]


Back to top | Article view | comp.soft-sys.math.mathematica


csiph-web