Page 2 - SUBHARADEV SEN AND AHMED
P. 2

SEN AND AFIFY

                           distribution  with probability density function (pdf) and cumulative
                           distribution function (cdf) as

                                               ,0        ;  ,     0                        (1)



                                         , 0       ,                                       (2)

                           respectively. On a note, Lai  and Mukherjee ([3], 1986) have
                           investigated some of the interesting properties of the density in (1) for
                           0 < p < 1 in reliability context.

                           For studying characteristics of multi-component system reliability, it
                           is rational to assume positive or negative interdependence for the life
                           lengths of the components. This interdependence might be caused due
                           to mechanical sticking of configurations, functional dependence on
                           common source of energy, incidental or environmental stresses, etc.
                           Although several works have been done in studying coherent
                           dependent systems for infinite  life lengths of components under
                           different configurations [see,  for example, Eryilmaz ([1], 2011),
                           Mukherjee and Sasmal ([5], 1977), Navarro et al. ([6], 2007), Navarro
                           and Rychlik ([7], 2007), Thomas ([9], 1986)],  no or little is yet
                           known about dependent systems with finite range lifetimes for even
                           two components.
                           The objective of the present investigation is to study the life behavior
                           of a two-component system under  bivariate finite range (BVFR)
                           distribution assumption. In this  article we introduce a bivariate
                           version of the finite range distribution given in (1) and thereby study
                           the detailed characteristics of a system life distribution in terms of
                           failure rate and reliability. The rest of the article is organized as
                           follows.

                           In section 2, BVFR is developed following the method of Farlie-
                           Gumbel-Morgenstern (FGM) copula and some basic properties useful
                           for the context are studied. Section 3 deals with distributions of
                           system lives and some basic distributional properties. Mean failure
                           times of dependent systems are compared, through numerical
                           considerations, with those of independent systems in subsection 3.1.
                           System reliabilities are studied  in section 4 along with numerical
                           tabulations. Section 5 deals with the failure rate functions of the

                                                           74
   1   2   3   4   5   6   7