On the selection of a best system from a group of systems with mutually dependent component lives

Abstract

This paper presents a probabilistic optimization problem of selecting the best system from a group of competing coherent systems with mutually dependent components by optimizing the probability that the system will meet a specified target lifetime and ranking the systems from best to worst according to the orders of those probability values from highest to lowest. The problem of comparing different system designs in search for the best configuration is of great concern to the reliability engineers, especially when the systems are of different orders and having mutually dependent components. Earlier works in this direction have discussed the problem for systems with independent components. The problem becomes complicated if the component lives are mutually dependent. The paper attempts to find an approach to solving the issue of comparing coherent systems of different orders with mutually dependent components. Here we select a system by minimizing the probability that the system will fail to meet a specified target lifetime, where the component lifetimes are mutually dependent, using a stochastic ordering criterion. A technique based on stochastic ordering has been proposed to compare systems of different orders. Some complex systems have been used to show that the result derived in this paper can be used effectively to make comparison of the systems with dependent components and to find the best system from a group of competing systems of different orders.