solver
c=CRITBINOM(n,p1,1-alpha)
so you only need to find the smallest n such that
binomdist(c,n,p2,true)<=beta
This is a problem that Solver can easily solve.
Jerry
"John" wrote:
Find c and smallest n that satisfy the constraints
binomdist(c,n,p1,true)=1-alpha
binomdist(c,n,p2,true)<=beta
For the case p1=.001, p2=.002, alpha=.05, beta=.05, solver found the
solution
c=23, n=16288
Solver could not find the solution for the case p1=.01, p2=.05,
alpha=.05, beta=.10.
The solution, found by trial and error, is c=3, n=110.
Is this an idiosyncrasy of solver, or am I doing something wrong?
John Adams
|