diff options
| author | tlatorre <tlatorre@uchicago.edu> | 2019-03-16 12:31:13 -0500 |
|---|---|---|
| committer | tlatorre <tlatorre@uchicago.edu> | 2019-03-16 12:31:13 -0500 |
| commit | e6a5af1eae49c773f205fb81f19fbce21f67c3f8 (patch) | |
| tree | 83142c9426aa6971c646de6bcaa00ac0543423e7 | |
| parent | a1edc8320b6fc219c03ee9433b630310bbdbd91c (diff) | |
| download | sddm-e6a5af1eae49c773f205fb81f19fbce21f67c3f8.tar.gz sddm-e6a5af1eae49c773f205fb81f19fbce21f67c3f8.tar.bz2 sddm-e6a5af1eae49c773f205fb81f19fbce21f67c3f8.zip | |
switch to using SBPLX for the minimization
Based on some initial testing it seems that the subplex minimization algorithm
performs *much* better than BOBYQA for multi-particle fits. It is also a bit
slower, so I will probably have to figure out how to speed things up.
| -rw-r--r-- | src/fit.c | 2 |
1 files changed, 1 insertions, 1 deletions
@@ -5252,7 +5252,7 @@ int fit_event2(event *ev, double *xopt, double *fmin, int *id, size_t n, double int status; /* Create the minimizer object. */ - opt = nlopt_create(NLOPT_LN_BOBYQA, 4+3*n); + opt = nlopt_create(NLOPT_LN_SBPLX, 4+3*n); nlopt_set_min_objective(opt,nlopt_nll2,&fpars); /* Guess the position and t0 of the event using the QUAD fitter. */ |