Benchmark methodology

How the replay works

The six rules, applied identically on both deposits:

• Monthly production timing: each annual period is spread across its twelve months by a precedence-aware topological placement, and a block is never placed before a predecessor.
• Drill-and-blast prep time: one month minimum gap on the directly-above edge. Every block is deferred forward until at least one month after the block directly above it was mined.
• Undermining stability guard: every block is deferred forward until the ±1 square on the bench directly above it has been mined.
• Active dig-face separation: two blocks closer than the configured separation distance are never mined in the same month.
• Fleet purchase capital: the fleet is sized once from the published solution's own busiest period sustained rate, never grown, and its capital is charged.
• Plant feed capacity: a month's mill feed never exceeds the published per-period processing limit divided across twelve months.

Discounting

The published solution is discounted in annual buckets, its own convention. The replay and the Cognomine run are discounted monthly.

What the replay cannot do

It inherits the published block selection and cannot re-choose what to mine, because the published formulation has no way to express fleet, haulage or drill-and-blast timing and so cannot be re-solved under them.

Sources

Both instances and their published solutions come from MineLib. Instance pages, solution files and retrieval dates are given on each benchmark page.