More packer nonsense

More packer nonsense

by ferris

This week was kinda slow as things were very busy at work, but then I was rewarded with a long weekend and things got MOVING.

Unfortunately most of the progress I made was on stuff I don't want to talk about, but it might be something for Solskogen, it's related to packers, and some other stuff I've written here before..

As for stuff I can actually talk about, continuing from the genetic algorithm stuff I did last week in order to improve my packer, I got some advice from rick about some of the details of the GA. Basically the main thing was that I was introducing a bit too much randomness (I was replacing 10 out of the 20 available values, which is kinda ridiculous), so I reduced that to only changing 2 of the values. I let stuff run again and ended up converging on slightly better models, now with no expansion in the executables from before I started this. Yay!

Yeah so not much I can write about publicly yet but that should change in the next week or two as I get stuff rolling a bit more :)