MathGroup Archive 2006

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

Search the Archive

principal value

  • To: mathgroup at smc.vnet.net
  • Subject: [mg67952] principal value
  • From: dimmechan at yahoo.com
  • Date: Thu, 13 Jul 2006 06:55:29 -0400 (EDT)
  • Sender: owner-wri-mathgroup at wolfram.com

Hi specialists.

I have to evaluate the following integral (from an elasticity project)

\!\(Integrate[ArcTan[x]\/\(x - p\), {p, 0, 1}, {x, 0, 1}]\).

Since p take values in the interval {0,1}, the option PrincipalValue
must be used.

\!\(Integrate[ArcTan[x]\/\(x - p\), {p, 0, 1}, {x, 0, 1},
PrincipalValue \
\[Rule] True]\)

\!\(\*FormBox[
  RowBox[{\(1\/32\), " ",
    RowBox[{"(",
      RowBox[{
        RowBox[{"32", " ",
          TagBox["C",

      Function[ {}, Catalan]]}], "+
        ", \(4\ \(\(log\^2\)(2)\)\), "-", \(Ï?\ \((Ï? + log(16))\)\)}],
")
          "}]}], TraditionalForm]\)

symb=N[%]

0.395399

I want also to compute previous numerically. To this end, I work as
follows:
First,

Needs["NumericalMath`"]
$Version
5.2 for Microsoft Windows (June 20, 2005)

\!\(f[x_, p_] := ArcTan[x]\/\(x - p\)\)

cauchy1[g_,x_,p_]:=CauchyPrincipalValue[g,{x,0,{p},1}]

\!\(\(mat1 =
     Table[{p,
       cauchy1[ArcTan[x]\/\(x - p\), x, p]}, {p, 0.0001, 0.9999,
0.0001}];\)\)

For p=0 the integral of f[x,p] in the interval 0<=x<=1 converges.

Integrate[f[x,0],{x,0,1}]

C

mat2=N[%]

0.915966

For p=1 the integral of f[x,p] in the interval 0<=x<=1 diverges.

Integrate[f[x,1],{x,0,1}]

\!\(\*FormBox[
  RowBox[{\(Integrate::"idiv"\), \(\(:\)\(\ \)\), "\<\"Integral of \
\\!\\(TraditionalForm\\`\\(\\(\\(
      tan\\^\\(-1\\)\\)\\)(x)\\)\\/\\(x - 1\\)\\) does not converge on
\
\\!\\(TraditionalForm\\`\\({0, 1}\\)\\).
\\!\\(\\*ButtonBox[\\\"Moreâ?¦\\\", \
ButtonStyle->\\\"RefGuideLinkText\\\", ButtonFrame->None, \
ButtonData:>\\\"Integrate::idiv\\\"]\\)\"\>"}], TraditionalForm]\)

\!\(TraditionalForm\`â?«\_0\%1\(\(\( tan\^\(-1\)\)(x)\)\/\(x -
            1\)\) \[DifferentialD]x\)

However (in the Hadamard sense)

Integrate[f[x,1],{x,0,1},GenerateConditions\[Rule]False]

\!\(\*FormBox[
  RowBox[{
    TagBox["C",
      Function[ {},
     Catalan]], "-", \(1\/8\ Ï?\ \(log(2)\)\)}], TraditionalForm]\)

mat3=N[%]

0.643767

So, I consruct a new list:

mat4=Insert[mat1,{0,mat2},1];
mat5=Insert[mat4,{1,mat3},Length[mat4+1]];

Then, I use the command to evaluate the integral

num=ListIntegrate[mat3, 2]

0.395833

The numerical result is different. What am I loosing something?

P.S. Evaluating an integral in the Hadamard sense, is (in most cases)
of no cost in elasticity theory. Usually, such "losing" of a
singularity, as before in p=1, corresponds to rigid body displacement.


  • Prev by Date: Re: Use color with ListPlot?
  • Next by Date: Reverse axis for LogLogListPlot
  • Previous by thread: Re: Gradient and Hessian matrix of cumulative normal ditribution
  • Next by thread: Re: principal value