Hi,

I would like to apply watson lemma on f(t) = 1/((t+1)*((t/a)-1)), with a a constant, and x tending to infinity.

The problem is that my integral I have to compute is not from 0, but from a....

My integral is : I = int (exp(-x*t)*(t^(λ))*f(t), t=a..∞), with λ>-1, x->∞.

I know that if a=0, and (taylor expansion) f(t) = ∑ b_n*(t^n), n=0..∞, then

I = ∑ b_n*Γ(λ+n+1)/(x^(λ+n+1)), n=0..∞.

Hello,

I need to create a new bernoulli distribution such that P(I=1) = q = 1-P(I=-1).

I did U:=Distribution(PDF=(t->piecewise(t=-1,1-q,t=1,q,0))); but Maple gives me back:

module() option Distribution, Continuous; export...   instead of:

module() option Distribution, Discrete; export...  since the distribution is discrete!!

Do you know a way to force U to be a discrete distribution, cause otherwise I get all moments=0.