Build parameter inputs for an MLE grid

Copies a complete SBT parameter list for every row of a grid and inserts the fixed, natural-scale grid coordinates on their model log scales. Unlike get_grid(), the grid layout is supplied directly and can therefore retain explicit cell identifiers and assessment-specific coordinate ordering.

Usage

make_mle_grid_parameters(
  grid,
  parameters,
  parameter_columns = c(h = "par_log_h", psi = "par_log_psi", m0 = "par_log_m0", m10 =
    "par_log_m10"),
  cell_column = "Cell",
  allow_new_parameters = TRUE
)

Arguments

grid

A non-empty data frame with one row per grid cell.

parameters

A named model parameter list.

parameter_columns

Named character vector mapping natural-scale grid columns to log-scale model parameter names.

cell_column

Name of the optional grid column containing unique, positive integer cell identifiers.

allow_new_parameters

Logical; allow a mapped grid coordinate to be added when it is absent from the template parameter list. This is needed for nominal coordinates such as M0 under the length-based mortality model.

Value

A named list of complete parameter inputs, one per grid row.

See also

Other grid workflows: grid_mcmc_to_tmbfit(), grid_to_tmbfit(), run_grid(), run_grid_mcmc(), sample_grid(), select_balanced_grid_iters(), summarise_mle_grid(), validate_mle_grid_state_records()

Examples

fit <- sbt_fit(sbt_example_data())
fit <- sbt_add_parameters(fit)
grid <- expand.grid(
  h = c(0.55, 0.63, 0.72, 0.80),
  psi = c(1.50, 1.75, 2.00),
  m0 = c(0.40, 0.45, 0.50),
  m10 = c(0.065, 0.085, 0.105)
)
grid$Cell <- seq_len(nrow(grid))
grid_parameters <- make_mle_grid_parameters(grid, fit$parameters)
length(grid_parameters)
#> [1] 108