Abstract
A very general model for the classical group testing problem is presented in which the only thing known about the population is the mean nonconforming rate p. It is shown that several procedures studied in the literature for more restrictive models, including the popular Dorfman procedure, can be extended to the general model without much affecting the number of tests required. In particular, one such extension leads to a procedure with some optimal properties for the general model.
Additional information
Notes on contributors
F. K. Hwang
Dr. Hwang is a Member of the Technical Staff.