Work that combines the work of Deep Neuroevolution and Hypernetworks in the perspective of evolving agent controllers able to adapt to changes to their train environment. Both techniques are very recent at the time of writing.
The results obtained with the limited time and resources during the semester were very promising, so additional experimentation is underway togheter with a revised version of the report that will be submitted for peer review.

Unpublished - COMING SOON
First Report