Researchers have demonstrated a way to drastically accelerate resource allocation solutions from large-scale optimization models - while still producing near-optimal results.