[BUG] Constraint residuals wrong/ dont make sense
Opened this issue · 0 comments
Describe the bug
Two constraint equations both seem to be satisfied, as indicated by a negative CC (negated to make it positive before being passed to the solver), but the residual of the first is negative and residual of the second is positive (positive residual should mean constraint satisfied?).
I have an output file that shows 12 of the inequality constraints have negative residuals but the CC’s are showing only one of the constraints are violated. We should probably review our constraint equations to ensure the CC’s and residuals make sense and that everything being passed to the solver and output is valid.
Metric | Bound | Residue (err) | CC |
---|---|---|---|
bmaxtf (4.62711) | bmxlim (14) | -9.3722 | -2.026 |
sig_tf_case (84138147) | sig_tf_case_max (750000000) | 665861853 | -7.91 |
Steps to reproduce
Run the IN.DAT and observe the values in the table above. (The file won't converge but will write out an OUT.DAT)
LT-INSPIRED-QSP-ISSUE.IN.DAT.txt
LT-INSPIRED-QSP-ISSUE.OUT.DAT.txt
Expected behaviour
Positive residual = satisfied? Therefore both should be positive, (or both should be negative)