Reliability optimization of series-parallel systems with K-mixed redundancy strategy

Highlights: • In this paper, the redundancy allocation problem (RAP) is revisited. • A new redundancy strategy called K-mixed is implemented on the system. • A new mathematical model is developed and solved by GA. • Results demonstrate the K-Mixed strategy outperforms all the previous strategies in RAP. -- Abstract: This paper revisits the redundancy allocation problem, a well-known problem in reliability optimization, and implements a novel redundancy strategy, called K-mixed, to improve system reliability. The K-mixed strategy is a general form of a mixed strategy. The mixed strategy is a combination of active and standby redundancy strategies which was introduced in 2014. Initially, the mathematical formulation for calculating the reliability of the K-mixed strategy is investigated and then its power and efficiency are evaluated against different test problems and a famous benchmark problem. In order to solve the proposed benchmark problem, an efficient genetic algorithm is developed and the results are compared with those reported elsewhere. It is found that the proposed K-mixed strategy results in higher reliability in most situations than would otherwise be achieved by its mixed counterpart. It is concluded that by using the K-mixed strategy in an optimization model, system designers have the opportunity to select the best strategy for each subsystem from among all the different strategies available to design systems with maximum reliability.