SddpSummary struct
#include <gtopt/planning_lp.hpp>
Summary of the last SDDP solve (populated by the SDDP solver).
Carries the final-iteration gap metrics so that write_out() can emit them as extra columns in solution.csv. For monolithic solves this struct is left at its default (all-zero / false) values.
Public variables
- bool converged
- True if any convergence criterion was met.
- double gap
- Final primary gap (UB−LB)/max(1,|UB|)
- double gap_change
- Final stationary gap-change metric.
- int iterations
- Number of training iterations completed.
- double lower_bound
- Final lower bound.
- double max_kappa
- Global max condition number across all solves.
- bool stationary_converged
- True if stationary criterion triggered convergence.
- bool statistical_converged
- True if CI-based statistical criterion triggered convergence.
- double upper_bound
- Final upper bound.