ABSTRACT
Let (XIr , X 2r , ... , X kr ), r == 1,2, ... be k-dimensional random sample vectors from a multivariate exponential (MVE) distribution whose survival function is given by
where Xi > 0, Ai > 0 (i == 1,2, ... , k) and Ao ~ O. This is a special version of the MVE distribution given by Marshall and Olkin (1967) under the assumption that a failure is caused by k + 1 types of Poisson shocks on a system containing k components. The MVE distribution is desirable when the assumption of independence among components is questionable or false. Plenty of literature is available about the MVE distribution because of its applications in life testing, reliability and other fields.
