[Date Index] [Thread Index] [Author Index]

Re: Simplify Expressions with Non commutative multiplication

• To: mathgroup at smc.vnet.net
• Subject: [mg60330] Re: Simplify Expressions with Non commutative multiplication
• From: Robert Schoefbeck <schoefbeck at hep.itp.tuwien.ac.at>
• Date: Tue, 13 Sep 2005 06:07:11 -0400 (EDT)
• References: <dfrgic$fpu$1@smc.vnet.net> <dfudi2$1m1$1@smc.vnet.net>
• Sender: owner-wri-mathgroup at wolfram.com

thanks for the suggestion,
but the form of the cancellation in my previous message was  just a 
special case, the general form being

myDot[P1___,M1,inv[M1+M2+...]P2___]
+
myDot[P1___,M2,inv[M1+M2+...]P2___]
+
myDot[P1___,M3,inv[M1+M2+...]P2___]
+
...

should be transformed to

myDot[P1,P2].

i had thought about programming a mySimplify working on luists but the 
combinatorics is quite messy and i wanted mathematica to help me in the 
first place, not being an obstacle. i do think that my demand is very
close to what mathematica can actually do, there should be a tiny trick 
to cast  simplify on noncommutative expressions.

robert schoefbeck

David Bailey wrote:
> Hello,
> 
> One problem with patterns of the form a___,something,b___ (or even more 
> complicated variants) is that they can work, but involve a combinatorial 
> explosion of possiblities - so things get hopelessly too slow when you 
> scale the code up to work on real problems.
> 
> A more efficient approach might be to treat the contents of your myDot 
> as a list and process it from left to right either recursively or even 
> indexing the terms explicitly using [[k]].
> 
> David Bailey
> http://www.dbaileyconsultancy.co.uk
>