Application of subset simulation methods to reliability benchmark problems

This paper presents the reliability analysis of three benchmark problems using three variants of Subset Simulation. The original version of Subset Simulation, SubSim/MCMC, employs a Markov chain Monte Carlo (MCMC) method to simulate samples conditional on intermediate failure events; it is a general method that is applicable to all the benchmark problems. SubSim/Splitting is a variant of Subset Simulation applicable to first-passage problems involving deterministic dynamical systems. It makes use of trajectory splitting for generating conditional samples. Another variant, SubSim/Hybrid, uses a combined MCMC/Splitting strategy and so it has the advantages of MCMC and splitting; it is applicable to uncertain and deterministic dynamical systems. Results show that all three Subset Simulation variants are effective in high-dimensional problems and that some computational efficiency can be gained by exploiting and incorporating system characteristics into the simulation procedure.