[MUD-Dev] Efficient AI?

[olag at ifi.uio.no]

Brian Lindahl wrote:

> For this sort of solution to AI, I think a neural net would work
> better than genetic algorithms, since genetic algorithms have
> usually significant overhead associated with them, and, therefore,
> could only be used in preprocessing the hash function.

Ah, indeed. I only meant to suggest having static hash functions as
the sole rationale for my proposal was execution speed. Maybe you
can hardcode neural networks and get similar speed, but you probably
would need more stateinformation? So it will eat more memory and
thereby reduce the efficency of the cache?

I don't know more than the most basic things about ANNs, but it is
my impression that the effective ones are rather specialized, so you
would need to find the right topology anyway? GAs could of course be
used for that... ;-)

Can you mix a state machine and neural networks efficently? E.g. use
the neural network to select states and then reconfigure the network
based on the selected state in order to get more control and higher
speed? Might work for a more ad-hoc carefully hand coded network.

> I've planned on using neural nets to implement a similiar scheme
> for locality-based reputations, see:

>   http://www.mudconnect.com/discuss/discuss.cgi?mode=MSG&area=adv_code&message=11022

Ah, that was an interesting idea! What are the locality based
reputations used for? NPC-trade and such? (e.g. the elves only trade
with "good" cities?)

Ola.
_______________________________________________
MUD-Dev mailing list
MUD-Dev at kanga.nu
https://kanga.nu/lists/listinfo/mud-dev