Entropy-driven Monte Carlo simulation method for approximating the survival signature of complex infrastructures

An active learning Kriging‐based multipoint sampling strategy for structural reliability analysis

“…(4) The proposed multi-add point strategy is applicable to parallelize the computation of AKMP method, which can effectively reduce the number of iterations of AKMP method and improve computational efficiency. (5) For the problem of high-dimensional functions, the combination of FELF learning function, NCC convergence criterion and multi-add point strategy can significantly reduce the computational cost. These three important features enable researchers to establish a balance between accuracy and computational cost while taking into account the requirements of real-world problems.…”

“…Traditional reliability analysis methods include the first-order reliability method (FORM), 1 the second-order reliability method (SORM), 2,3 Monte Carlo simulation (MCS), 4,5 subset simulation (SS), 6,7 important sampling (IS), 8,9 directional sampling (DS), 10,11 etc. Among them, the MCS method has a simple calculation process but requires a large number of simulation calls to obtain high precision failure probability.…”

Sampling and active learning methods for network reliability estimation using K-terminal spanning tree

“…While this method reduces memory requirements by avoiding the evaluation of structure functions for all survival signature entries, its accuracy depends on the number of MCS samples, limiting its applicability to larger and more complex networks. Furthermore, Di Maio et al [21] develops an entropy-based MCS method to improve computational efficiency by using entropy to guide sampling in unknown survival signature regions. However, this method may face challenges when applied to networks with a larger number of unreliable nodes.…”

Ensemble of Artificial Neural Networks for Approximating the Survival Signature of Critical Infrastructures

“…Recently, the survival signature has been introduced in reliability assessment [6] for modelling dependencies [4], [7], [8], handling imprecise probabilities in a Bayesian scheme [9]- [11] and dealing with the complexity of the CIs [12]- [15]. The survival signature allows decoupling the structural information of the system from the probabilistic one [6], [16], which can be a major advantage for the reliability assessment of large systems.…”

“…Monte Carlo Simulation (MCS) combined with percolation theory [18] has been used to obtain an approximation of the survival signature for large systems. Entropy has been used to guide MCS in a way to efficiently allocate the simulation efforts over the whole survival signature [12]. Nonetheless, the search for every combination of functioning components of a large system makes the approximation of the survival signature computationally prohibitive.…”