This is a closed-loop simulation, not GP training. The GP/EM object is frozen. No heights are automatically adopted or retuned by scenario. The summary reports terminal true B/B_R, not EM input, and brackets for later refinement.
Usage
run_plateau_search(
om,
em,
heights,
parameters,
reference_years,
scenarios,
iterations,
years,
targets,
scenario_weights,
seed = 271828L,
cores = 1L,
checkpoint_dir = NULL,
resume = TRUE,
max_steps = Inf,
progress = NULL
)Arguments
- om
Operating model configured to save
B_over_BR.- em
One calibration EM.
- heights, parameters, reference_years
Arguments to
mse_plateau_design().- scenarios, iterations, years, seed, cores, checkpoint_dir, resume, max_steps, progress
Arguments to
run_mse().- targets
Positive terminal true B/B_R targets.
- scenario_weights
Named positive weights, one per scenario, summing to one.