Major optimality tolerance

This specifies the final accuracy of the dual variables.  On successful termination, SNOPT will have computed a solution (x, s, p) such that

 
where gap_j is an estimate of the complementarity gap for variable j (j =1:n+m).  The gaps are computed from the final QP solution using the reduced gradients d_j = g_j -pTa_j (where g_j is the jth component of the objective gradient, a_j is the associated column of the constraint matrix (A - I), and p is the set of QP dual variables):

In the major iteration log, maxgap appears as the quantity labeled "Optimal".