no_recursion

Do not recur over admissible DAGs for a given model ensemble

Specification

  • Alias: None

  • Arguments: None

  • JSON argument key: None

  • Default: true

Description

This (default) option supports restriction to the peer DAG (all control variates target the same truth model) in the ACV approach (approximate_control_variate and [GGEJ20]) and restriction to a hierarchical DAG (each approximation node points to the next approximation of higher fidelity/resolution, ending with the truth model at the root node) for weighted MLMC (weighted) and MFMC (multifidelity_sampling).

Note that the omission of the search_model_graphs block defaults to no searching over either DAGs or model subsets. The no_recursion option exists in order to enable search over the latter while suppressing search over the former, as shown below.

Examples

method,
  model_pointer = 'ENSEMBLE'
  approximate_control_variate acv_mf
    pilot_samples = 25 seed = 8674132
    search_model_graphs
      no_recursion
      model_selection
    max_function_evaluations = 500
{
  "method": [
    {
      "approximate_control_variate": {
        "model_pointer": "ENSEMBLE",
        "pilot_samples": [25],
        "seed_sequence": [8674132],
        "search_model_graphs": {
          "no_recursion": {},
          "model_selection": {}
        },
        "max_function_evaluations": 500
      }
    }
  ]
}