Solve a model in-process with HiGHS — solve

Uses the solver-object interface rather than highs::highs_solve(), because only that path returns dual values (constraint marginals) as well as primal levels.

Usage

solve_highs(model, lp = NULL, attach = TRUE, control = list(), verbose = FALSE)

Arguments

model: A multimod model with data attached.
lp: Optional pre-built LP from model_to_lp().
attach: Logical; attach primal values back onto model$variables as a long data.frame with named index columns, matching what read_csv_results() produces so downstream readers keep working.
control: Optional list of HiGHS options.
verbose: Logical; report progress.

Value

A list with objective, status, status_message, primal (a data.table of j, symbol, index tuple and value), dual (per row), the lp, and - when attach is TRUE - the updated model.

Examples

if (requireNamespace("highs", quietly = TRUE)) {
  r <- solve_highs(example_models$toy, attach = FALSE)
  r$objective
}
#> [1] 198000