matflow.tests.subset_simulation.generate_next_level_samples_ACS#

matflow.tests.subset_simulation.generate_next_level_samples_ACS(num_chains, num_states, dimension, chain_seeds, chain_g, all_x, all_g, level_idx, master_seed, target_pf, threshold, chains_per_update, prop_std=1.0, lambda_=1.0)#

Adaptive conditional sampling algorithm for generating states in the subset simulation level (aka adaptive subset infinity).

Parameters:

  • chains_per_update – The number of Markov chains (as a fraction of the number of samples per level) that will run before lambda_ is updated. As an integer number, known as Na elsewhere.

  • prop_std – Initial variance of the proposal distribution.

  • lambda – Initial scaling parameter.