0 Membres et 1 Invité sur ce sujet
My review of the 10+ cycle results has shown that the simulator seems to settle into a couple of states over long runs. Firstly, there are tiny parasites who successfully execute hundreds of thousands of cycles (all of these from other creatures code), and secondly the vast majority of creatures live just long enough to reproduce once or twice but then die out. There seems to be a dominant creature strand which is not quite as long as the original parent code, but is still obviously successful in reproducing.I have also found that one particular creature strand (when it has been included in a simulation) survives better than all the others. This is rather interesting as it is one of the simpler creatures and would suggest that conditions at the beginning of the simulation tend to favour it over the more complicated (but ultimately more efficient) creatures.Some of the things I'll be looking at in the near future are creature design, adjusting the starting conditions to ensure a level playing field for all creatures (maybe by delaying mutations for a few thousand instructions) and the criteria for killing creatures during mass extinctions and reaper events (because I have a feeling that these currently favour parasites).
http://www.boinc-af.org/content/category/11/68/219/
D'un côté, le portail de L'AF est à l'abandon depuis 6 mois.
cette page est loin d'être à jour visiblement, il doit manquer une bonne dizaine de projets
ça roule ! Va falloir mettre de la puissance ... Ars Technica semble très forte !
Et encore il n'y a pas beaucoup de monde