MathGroup Archive 2008

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

Search the Archive

Re: D

  • To: mathgroup at smc.vnet.net
  • Subject: [mg94553] Re: D
  • From: dh <dh at metrohm.com>
  • Date: Tue, 16 Dec 2008 02:34:08 -0500 (EST)
  • References: <ghtjb8$r6c$1@smc.vnet.net>


Hi Cetin,

try not to overload your question with insignificant details. The epsj 

are unimportant for the derivatives, I do not write them. Further, I 

give the principle for only 2 functions and variables, otherwise it will 

be a mess to check if everything is correct.

We want the Jacobian of f1[y1,y2] and f2[y1,y2]. This is sort of an 

outer product between fi and derivative in direction yj:

funs={f1,f2};

vars={y1,y2}

Outer[D[#1@@vars,#2]&,funs,vars]

For many dimensions it mqay be profitable to generalise this further. If 

we denote the functions f1,.. by f[1],.. and the variables y1,.. by y[1] 

we can write:

funs=Array[f,2];

vars=Array[y,2];

Outer[D[#1@@vars,#2]&,funs,vars]

For 21 dimensions, you simply have to reaplce 2 by 21.

hope thsi helps, Daniel



Cetin Haftaoglu wrote:

> Dear mathematica users,

> 

> 

> 

> I have a to  evaluate the Jacobian matrix from  a vector of function fi,

> i=1,21 with the inner variables yi, i=1, 21 and epsj j=1,6.

> 

> So I have to evaluate D[f_i [yi,epsj] , y_j], i= 1, 21, j=1, 21. How can I

> write this expression in a short form, without to say that the functions f_i

> depent on the inner

> variables yi. When I write D[f_i,yi] I receive 0 because mathematica thinks

> that I want derive a Constant. When I say

> 

> 

> 

> D[f_i[y1,y2,y3,=85.,y21,eps1,=85eps6],y_j] ], i= 1, 21, j=1, 21

> 

> 

> 

> I get the right expression but a very great expression (output). How can I

> reduce this output?

> 

> 

> 

> TIA,

> 

> Regards,

> 

> 

> 

> Cetin Haftaoglu

> 

> Arbeitsgruppe " Modellierung und Simulation in der Werkstoffmechanik"

> 

> BAM =96 Bundesanstalt f=FCr Materialforschung und -pr=FCfung

> Fachgruppe V.2 =96Werkstoffmechanik

> 

> 

> 

> Unter den Eichen 87

> 

> D-12205 Berlin

> 

> Deutschland

> 

> 

> 

> Tel:    +49-30-8104-3194

> 

> Fax:   +49-30-8104-1527

> 

> Email:  <mailto:cetin.haftaoglu at bam.de> cetin.haftaoglu at bam.de

> 

> Web:   www.bam.de




  • Prev by Date: Re: Evaluate ?
  • Next by Date: Re: Recognising parameters in function
  • Previous by thread: D
  • Next by thread: branch of (-1)^(1/3)