Use a 2-stage approach to sextupole optimizations. Start with optimized sextupole distribution from 2010.05.04 (best to date), which was derived from 2fam. Apply new constraints from 2010.06.10 with reduced weights on some parameters to this optimized distribution.
Next, try starting with 2fam distribution. Reduce weight on all "second-order" terms (second-order chromaticities, alpha_xx,xy,yy) by 4-5 orders of magnitude. Optimize. Then apply constraint file from 2010.05.04 that yielded best results to date.
Try starting with 2fam distribution. Reduce weight on all "second-order" terms (second-order chromaticities, alpha_xx,xy,yy) by 4-5 orders of magnitude. Optimize. Then apply constraint file from 2010.06.10.