Skip to content
Prev 179012 / 398506 Next

integrate with large parameters

On 01/05/2009 5:46 AM, Andreas Wittmann wrote:
I think if you can work out the mode of the function, and divide by 
that, it should be possible to make things work.  (The main problem will 
be that a large range of function values will underflow to zero. 
Hopefully those have a negligible integral, but if integrate evaluates 
the target there too often, it might think the integral is zero.)

Duncan Murdoch