Abstract
System structure optimization is a well-studied problem in the field of reliability engineering, which aims to achieve the best possible reliability versus cost solutions for the system design. This problem is usually solved for systems that do not change their task and configuration during the mission. However, many practical systems are phased-mission systems (PMS), where the mission involves multiple, consecutive, and nonoverlapping phases of operation. An accurate analysis of PMS must consider the dynamics in system configuration, success criteria, and element behavior as well as statistical dependence of element states across phases. In this paper, we propose a method for solving the structure optimization problem of multistate PMS consisting of nonidentical nonrepairable binary elements. The system configuration and demand as well as the failure distributions of system elements can change from phase to phase. The proposed approach is based on a recursive algorithm for the reliability evaluation of PMS and a genetic algorithm for the structure optimization of PMS. The method is illustrated using an example of a six-phased mission running on an airborne distributed computing system.