A multi-objective multi-state series-parallel redundancy allocation model using tuned meta-heuristic algorithms

Document Type


Publication Date





In this paper, we study a series-parallel multi-objective multi-state redundancy allocation problem (MSRAP) with known performance levels and corresponding state probabilities. The problem is comprised of multiple subsystems in series and each subsystem is comprised of multiple components in parallel. The system components have a range of performance level from complete working to complete failure. The subsystems contain homogenous redundant components and the component prices come under an all-unit discount policy if a unique brand (type) is chosen for purchasing all subsystem components. Each component is characterised by its cost, weight and availability. The goals are to find the optimal combination of the components in each subsystem that maximises system availability and minimises the total cost under a weight constraint. We propose a multi-objective harmony search (MOHS) algorithm, a non-dominated sorting genetic algorithm (NSGA-II), and a multi-objective genetic algorithm (MOGA) to solve this problem. In addition, the Taguchi method is utilised to tune the parameters in each algorithm. We use a number of numerical examples to demonstrate the applicability and exhibit the efficacy of the three algorithms. The results show that the MOHS outperforms the NSGA-II and MOGA with respect to all of the considered metrics.




Di Caprio, D., Santos-Arteaga, F.J. and Tavana, M. (2017) ‘A Novel Decision Support Framework for Computing Expected Utilities from Linguistic Evaluations,’ International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems, Vol. 25, No. 6, pp. 1005-1018.