[MUD-Dev] Efficient AI?

[Brian Lindahl]

With reguard to generating hash functions, I stumbled upon this the
other day: http://ftp.cnr.it/gnu/gperf/. It may help. Now, with
reguard to your respose:

olag at ifi.uio.no wrote:

> 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?

Generally, neural networks CAN be hard-coded, but this removes any
ability to train them in an online fashion (help generate better AI
over time). Also, this would essentially be a genetic algorithm,
using a large training set to generate a neural network that
provides a reliable and believable output. Neural networks are best
used to solve dynamic problems, but CAN be used statically.

> would need to find the right topology anyway? GAs could of course
> be used for that... ;-)

Possibly. However, as I mentioned, I have only planned on using them
and haven't acquired enough experience using them yet to comment on
this. There may be a good way to generate a topology that works
without having to experiment a lot.

> 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?

For this, I would probably generate a neural network per state. You
could dump the results into a database and use them to train the
networks in an offline fashion.

> 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?)

All NPC interactions. Anywhere from bumping into one in a busy
marketplace to common jobs to poltical alliegences.
_______________________________________________
MUD-Dev mailing list
MUD-Dev at kanga.nu
https://kanga.nu/lists/listinfo/mud-dev