42 USA Evaluation #2

Oh, and logic starts with F1. Each command arrow indicates the movement of the arrow on the board, the F1-F3 commands call their respective functions. And coloring squares in the function area via color commands means those commands are executed when the arrow on the board is present over a matching color square.

Also, FYI the Evaluation is 2 hours so I got to this problem in the latter half of the evaluation. Don't bother with trying to solve it on the site to help me if you want to reply in time. Just try solving it with the image.

Bump 2.

Use q learning

>>60134769
q learning?

>>60133908
What happens when it goes on a gray tile?

>>60134872
You mean the boundary of the board? It fails the test and it starts over.

>>60134885
What do the colors on the function tiles do?

>>60134900
Like I mentioned, it makes it so it only executes when the arrow on the board is on the matching color tile. 42 tells you nothing and you essentially have to figure everything out yourself.

>>60134938
>42 tells you nothing and you essentially have to figure everything out yourself
Yeah I can't very well run that fucking image tho. I don't see a way to do it without either memory and comparisons or knowing about delays with execution on the commands.

>>60135062
And that's why it stumped me. Just simple branch logic would make it so much easier.

>>60135084
Also, technically there is a stack and for every function call, it pushes the rest of the commands in the function to the end of the stack, but there's no easy way to pop except not calling a function at some point and letting execution end, which when that happens means that you either finish collecting the stars with the actions left in the stack or you lose the problem and have to start over.

>>60135126
But the thing is only one of the problems I encountered utilized the stack queue to finish up a problem, and for this problem it seems like there is too many tiles of traversal to account for in leftover commands.