A swarm intelligence-based approach for reliability–redundancy allocation problems

Abstract

This paper presents a new algorithm belonging to the class of swarm intelligence methods, called the adaptive simplified PSO (ASPSO)-based algorithm, for solving reliability–redundancy allocation problems. In this constrained nonlinear mixed-integer problem, both the number of redundant components and their reliability in each subsystem are to be decided simultaneously so as to maximize the reliability of the system. The proposed ASPSO operates with a new updating model to adjust the position of particles, without dealing with velocity. In addition, a randomization technique, based on the dispersion of particle bests through the search space, is used to speed up the convergence of the proposed approach and prevent it from being trapped within the local optimum. Moreover, to control the balance between exploration and exploitation, during the search process, two adaptive functions are utilized. The simulation results of four different benchmarks for the reliability–redundancy allocation problem are reported and compared. Accordingly, the solutions given by the new presented approach are all superior to those best known solutions provided by several methods in the literature.