> To be exact every participant has a maximum of 1024 tokens at his disposal to craft the best chess bot they can.
I'd be quite tempted to try linear genetic programming with a variant of a Brainfuck-style UTM for this kind of constraint. Assuming 1024 tokens = 1024 bytes = 1024 instructions, I think there is some degree of performance that is feasible.
This is really interesting to me because hand-coding BF to do something like chess is absolutely beyond me. I wouldn't even know where to begin. A LGP-coded BF program would almost certainly outperform anything I could design with my own hands.
I'd be quite tempted to try linear genetic programming with a variant of a Brainfuck-style UTM for this kind of constraint. Assuming 1024 tokens = 1024 bytes = 1024 instructions, I think there is some degree of performance that is feasible.
This is really interesting to me because hand-coding BF to do something like chess is absolutely beyond me. I wouldn't even know where to begin. A LGP-coded BF program would almost certainly outperform anything I could design with my own hands.