why getting odd numbers of designs during the heuristic process? - Githubissues

drbenvincent / darc_toolbox

Run adaptive decision making experiments

MIT License

16 stars 2 forks source link

why getting odd numbers of designs during the heuristic process? #19

Closed drbenvincent closed 5 years ago

drbenvincent commented 5 years ago

I'm getting weird logging information about the number of designs throughout the heuristic process

INFO:root:Getting design for trial 4
DEBUG:root:2052 designs initially
DEBUG:root:2028 designs after removing prior designs
DEBUG:root:2704 designs after removing highly predicted designs

The removal of prior designs should only remove a number equal to the number of trials run.
We should not end up with more designs after removal of highly predicted designs!