MathGroup Archive 2004

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

Search the Archive

RE: Alternative to defining 'operator' function?

  • To: mathgroup at smc.vnet.net
  • Subject: [mg47565] RE: [mg47505] Alternative to defining 'operator' function?
  • From: "Owen, HL (Hywel)" <H.L.Owen at dl.ac.uk>
  • Date: Fri, 16 Apr 2004 05:20:34 -0400 (EDT)
  • Sender: owner-wri-mathgroup at wolfram.com

Thanks to David, Bob and Paul for useful suggestions.

Rest[FoldList[#2.#1&, v, {R1, R2, R3}]] and
Rest@FoldList[Dot[#2, #1] &, v, {R1, R2, R3, R4}]

are basically equivalent and work for an arbitrary list of matrices -
thanks! Couldn't see for looking... Looks like it'll work for any function
of two variables I guess?

Paul's idea is a good one if you need to generate a symbolic set of
expressions (not right for my problem though):

Subscript[R, n_][v_] := Subscript[R, n]
Rest[ComposeList[Subscript[R,1], Subscript[R,2], Subscript[R,3]}, v]]

Hywel

> What about...
> 
> Rest@FoldList[Dot[#2, #1] &, v, {R1, R2, R3, R4}]
> {R1.v, R2.R1.v, R3.R2.R1.v, R4.R3.R2.R1.v}
> 
> David Park
> djmp at earthlink.net
> http://home.earthlink.net/~djmp/ 
> 
> From: Owen, HL (Hywel) [mailto:H.L.Owen at dl.ac.uk]
To: mathgroup at smc.vnet.net
> 
> Hi folks,
> 
> I often have programming problem where I'd like to calculate 
> a set of dot
> products, e.g. applying a list of square matrices 
> {R1,R2,R3...} to a vector
> v to obtain:
> 
> {R1.v,R2.R1.v,R3.R2.R1.v,...}
> 
> or other functions like that.
> 
> The method I've been using is to define an 'operator' function, e.g.
> 
> DotOperator[M_] := Dot[M, #] &
> 
> Then we have:
> 
> In: DotOperator[R][v]
> Out: R.v
> 
> as wanted, so that we can define a ComposeList as
> 
> In: Rest[ComposeList[DotOperator[#] & /@ {R1, R2, R3}, v]]
> Out: {R1.v, R2.R1.v, R3.R2.R1.v}
> 
> to obtain the result we want.
> 
> Is there a simpler way than this that doesn't involve 
> defining functions
> like DotOperator?
> 
> Thanks,
> 
> Hywel
> 
> 
> 


  • Prev by Date: Re: Solvability sensitive to small changes in numerical input
  • Next by Date: RE: Re: DirectMath experience
  • Previous by thread: RE: Alternative to defining 'operator' function?
  • Next by thread: Re: Alternative to defining 'operator' function?