performanceProfile

Performance-profile data across an experiment's traced runs: for each solver case and each budget fraction on a grid, the fraction of (problem, macro-rep) cells whose best penalized objective reached the problem's gap basis plus tau within that fraction of the replication budget.

Requires captured iteration traces and problems with a gap basis; runs without either are excluded. The solve test compares the (minimization-oriented) penalized objective against the oriented basis plus tau, which equals the raw objective comparison for problems without response constraints.

Parameters

Performance-profile data pooled across the blocks of one study (see the pooled mcbDataMap). The blocks must share a replication budget, since the profile's x-axis is the fraction of that budget; pooling experiments with different budgets would put incomparable cells on one axis.

Each problem needs one solve threshold across the whole pooled set. Problems measured against a reference solution already carry the same basis in every block and it is used directly. A problem with no reference is gapped against the best objective found within its experiment, so the blocks disagree — the basis is then recomputed as the best across every pooled run of that problem, which is the value a single unblocked run would have recorded. With one experiment there is nothing to disagree with, so a single-experiment profile is unchanged.

Parameters