> > It seems that polynomials with even power in the
> > leading monomial produce some issues. However
> > all of these failures have very reasonable results.
> > Maybe they are out of the scope of Risch for some
> > reason?
> 
> Well, at first I was surprised how much FriCAS can do... 

> The 'exp(-x^2)' function is a "canonical" example of function
> having no elementary integral.

Yes, right.

> Using integration by parts it is easy to see that integral
> of 'x^(2*k)*exp(-x^2)' is a constant times integral of
> 'exp(-x^2)' plus elementary part, so in this case also
> there is no elementary integral.
>
> In case of 'x^(2*k+1)*exp(-x^2)' using integration by parts
> one can easily find elementary integral.  So Risch algorithm
> handles odd powers with no problems, while even powers have
> no elemetary integral so Risch gives up.

I see, with odd exponent we hit the base case int(x*exp(-x^2))
which we can solve while for even exponent we find int(exp(-x^2)).
This is fine so far concerning the algorithmic part.

> It is interesting that Risch still can do orthogonal cases (with
> proper weight). AFAICS this is because nonelementary parts cancel.

Hmm, do they?
The polynomial H_m*H_n has either only odd or only even monomials.
Odd cases should be no problem at all, as we found above.

> In other words we have sum where each term separately have no
> elementary integral but the sum does.

>  In m = n case the main contribution comes from nonelementary
> part and Risch can not do this.

> In case m = n = 0 the integral is
> handled by pattern matcher, but the case m = n > 0 is not
> handled by pattern matcher.

Ah well. In that case we have a polynomial times the exponential.
Maybe one could extend the pattern matcher for this case, maybe not.
(Another Idea would be to use the Meijer-G Ansatz to integration.)

Thanks for this explanation. I see now why I get the results I got.
I'd have to work out all the cases on paper now for more insight.

-- 
You received this message because you are subscribed to the Google Groups 
"FriCAS - computer algebra system" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/fricas-devel?hl=en.

Reply via email to