Why a solve stopped, read off its last convexification: model, the
model's predicted decrease under ISchedule.improved or within its own
accuracy (IStep.accuracy); step, a step taken under
ISchedule.least, or a rejected proposal under it at an already feasible
iterate (numerical no progress, not a stationarity certificate);
collapsed, no step taken, the region fallen under
ISchedule.least or the tries spent; rounds, the convexifications ran
out (IOptions.rounds) with steps still being taken; infeasible, a
steered iterate at which the step of the violation alone, inside its
region, reduces the violation by no more than IOptions.stationary
while it still violates by more: a stationary point of the violation.
Why a solve stopped, read off its last convexification:
model, the model's predicted decrease underISchedule.improvedor within its own accuracy (IStep.accuracy);step, a step taken underISchedule.least, or a rejected proposal under it at an already feasible iterate (numerical no progress, not a stationarity certificate);collapsed, no step taken, the region fallen underISchedule.leastor the tries spent;rounds, the convexifications ran out (IOptions.rounds) with steps still being taken;infeasible, a steered iterate at which the step of the violation alone, inside its region, reduces the violation by no more thanIOptions.stationarywhile it still violates by more: a stationary point of the violation.