MathGroup Archive 2002

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

Search the Archive

Re: UnitStep in Mathematica 4.1

  • To: mathgroup at smc.vnet.net
  • Subject: [mg38210] Re: [mg38180] UnitStep in Mathematica 4.1
  • From: Jack Goldberg <jackgold at umich.edu>
  • Date: Thu, 5 Dec 2002 03:31:04 -0500 (EST)
  • Sender: owner-wri-mathgroup at wolfram.com

Hi Matt,

Every so often a user like yourself asks a question involving, directly or
indirectly, about UnitStep.  Becasue I too found Mathematica's implementation of
UnitStep "spotty" I wrote for myself a package that handles all (?)
reasonable cases and a few "unreasonable" ones.  I keep meaning to submit
it for others to use but one thing or another keeps me from it.  Oh
well...

Now to answer your question.  All functions with jump continuities can be
written as a sum of the form

	f0[x] + f1[x]*UnitStep[x-b1] + f2[x]*UnitStep[x-b2]
		+...+ fn[x]UnitStep[x-bn] +g[x]

where each fi[x] has no jumps and g[x] is a similar sum of DiscreteDelta
functions.  But when one integrates, the integrals of DiscreteDelta  is
zero so the g[x] term can be ignored.  Now the critical point:  Mathematica knows
how to integrate sums of the type given above!  So all you have to do is
write  UnitStep[-x]*UnitStep[x]  as a sum of this type.  Since 0 is the
only jump of this function (and neglecting g[x]) we have

UnitStep[-x]*UnitStep[x] = 1 + f1[x] UnitStep[x]

But at x=0, the lhs = 1 so  f1[x] = 0 and you have your answer.

My program does all this simplification automatically.

Jack (Dr. UnitStep) Goldberg


On Wed, 4 Dec 2002, Matt Pillsbury wrote:

> Hi,
>
> I have been relying on the UnitStep function in a project, and in
> general it seems to work pretty nicely. However, there is one simple
> problem that I'm having with it. Mathematica doesn't know that
>
> Integrate[UnitStep[-x] UnitStep[x], {x,-1,1}]
>
> should be zero. Is this because the integrand is 0 everywhere but at
> the origin, where it is 1, and Mathematica doesn't know what to do
> with the contribution from a set of measure zero? Or am I doing
> something else wrong?
>
> And is there a simple fix for my problem?
>
> Thanks,
> Matt Pillsbury
>



  • Prev by Date: Re: 3D Animations are killing my system
  • Next by Date: Re: 1^Infinity
  • Previous by thread: Re: UnitStep in Mathematica 4.1
  • Next by thread: Re: Stereograms in Mathematica