#75 Convolution problem

[Manfred Kremer]

@Rainer and Arthur: Thank you very much for fixing the integration bugs.
Unfortunately today I wanted to calculate a series of convolutions, but the results were not satisfactory. Is it a bug or did I do anything wrong?

[Rainer Schöpf]

Hi Manfred,

it's actually both: first, there is a bug in the display routines. You see that if you switch to text output with "off fancy;".

Your input doesn't do what you expect it to do: an assignment to an operator defines the exact left hand side, not a general function. I.e.

f1(x) := x*x;

tells Reduce what f1(x) means, but nothing about, e.g., f1(y).

What you want can be achieved by either a rule like

let f1(~x) => a*int((1-exp(-a*(x-y)))*exp(-a*y),y,0,x);

or a procedure like

procedure f1(x); a*int((1-exp(-a*(x-y)))*exp(-a*y),y,0,x);

With both methods you get the desired result.

Rainer

[Rainer Schöpf]

I committed a fix to the display problem.

Rainer