Difference between revisions of "DOCK GA Development Goals"
From Rizzo_Lab
(→To Do List) |
(→To Do List) |
||
| Line 32: | Line 32: | ||
##pick location of mutation based on something | ##pick location of mutation based on something | ||
##pick mutation type based on behavior of ensemble | ##pick mutation type based on behavior of ensemble | ||
| − | ##too large boost deletion | + | ##molecules too large boost deletion |
| − | ##too small more | + | ##molecules too small, add more groups |
##change ...boost replace/sub | ##change ...boost replace/sub | ||
##mutation type selection based on probability vs ensemble | ##mutation type selection based on probability vs ensemble | ||
Revision as of 10:40, 13 May 2019
| Tasks | src | Owner | Complete? |
|---|---|---|---|
| Fix bug that collapsed atom coordinates | everywhere? nowhere? | everyone | No |
| Added Delimeter header | conf_gen_ga | LEP | Yes |
| Fix xover only feature | conf_gen_ga | LEP | Yes |
| Put in error messages for mut_rate > 1 | conf_gen_ga | LEP | Yes |
| Manual user-defined mutation type | conf_gen_ga | LEP | Yes |
| Remove check only option | conf_gen_ga | LEP | Yes |
To Do List
- -xover (guided based on score) - Good v Good ; Bad v Good ; Bad v Bad
- nonexhaustive xover (pick subset of xover based on probability)
- 2-3 point xover at once
- adaptive maintenance ensemble based on ensemble convergence
- bring in new parents based on convergence
- Mutations-
- adaptive mutation rate
- pick location of mutation based on something
- pick mutation type based on behavior of ensemble
- molecules too large boost deletion
- molecules too small, add more groups
- change ...boost replace/sub
- mutation type selection based on probability vs ensemble
- complete x # y mutation so far so less prevalent etc
- 3 layer subs do no work so don't do them
- replace > 1 segment
- fitness-
- turn on and off niching adaptive/extinction
- reduce boost of fragments and all poor mols with niching
- pareto/mulitobjective ga
- selection-
- metropolis selection for tournament/roulette
- adaptive keep #p and #o
- extinction-
- automated vs on convergence
- stop-
- convergence
- which molecules are best-
- best first pruning - now uses descriptor score even if niching ned to delta to fitness/niching when used
