MathGroup Archive 2008

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

Search the Archive

Re: Re: delta function

  • To: mathgroup at
  • Subject: [mg89950] Re: [mg89909] Re: delta function
  • From: Daniel Lichtblau <danl at>
  • Date: Wed, 25 Jun 2008 06:25:29 -0400 (EDT)
  • References: <g3ihpo$fdt$> <>

Magician wrote:
> On Jun 21, 4:31 am, Magician <jadoo.d... at> wrote:
>> I am integrating over a function (not written in mathematica syntax)
>> u=F(x)   e^(-  (x-xo)^2/t  ) /t  ,
>> i know in the limit t ->0,  e^(-  (x-xo)^2/t  ) /t = delta(x-xo), but
>> how do i get mathematica to recognize this.
>> In mathematica, how can i construct hings like the Sokhotskyi-Plemelj
>> formula ?
> hey somebody plz comment????

Here are a few comments.

(1) Use Mathematica syntax to denote Mathematica expressions.
(2) Use correct formulations. Your denominator is not correct, if what 
you seek is the effect of a delta function.
(3) Don't throw in things that are irrelevant to your message. In this 
case, u=F(x) is irrelevant to your message.

About that limit...yes, it might be useful to have it give a DiracDelta 
result. At present what one can do, in integrating a concrete function, 
is use Limit outside Integrate to get the effect of integration against 
a delta function.


ee = Exp[-(x-xo)^2/t]/t^(1/2);

In[29]:= InputForm[Limit[Integrate[ee*x^2, {x,-Infinity,Infinity}],
          t->0, Assumptions->Element[{x,xo},Reals]]]
Out[29]//InputForm= Sqrt[Pi]*xo^2

This has the advantage, moreover, of being mathematically correct 
whereas having Limit return a DiracDelta might go against some standard 
Limit behavior (I have not fully thought this out).

Daniel Lichtblau
Wolfram Research

  • Prev by Date: Re: How can I evaluate this statement?
  • Next by Date: Piecewise and multiple values
  • Previous by thread: Re: delta function
  • Next by thread: Re: delta function