[MUD-Dev] NPC AI
Mike Sellers
mike at online-alchemy.com
Sat Oct 18 17:27:46 CEST 1997
At 10:57 PM 10/11/97 PST8PDT, Travis Casey wrote:
>Brian Price <blprice at bedford.net> wrote:
>>As a side note, I have read of solutions that used a mixture of
>>neural nets and fuzzy logic where neural nets provided some learning
>>capability. Unfortunately I don't have enough expertise with AI to
>>evaluate the suitability of such a combination for muds. Indeed I
>>don't have the necessary expertise to evaluate the combination of
>>fuzzy logic with genetic algorithms. I'm hoping that the AI gurus in
>>this group can provide an expert opinion.
>
>Expert networks may be what you're referring to here -- I'm not really
>familiar enough with the field to know for sure. The only reason I've
>ever heard of them is because they were the field of research of a
>CS professor I worked for.
The original poster may be referring to Kohonen nets here. These are a
powerful Bayesian classifier that are also easy and compact to program.
Being a type of Bayesian classifier, they make excellent use of "fuzzy
logic", and with a few tweaks on the basic algorithms they can quickly
generate an optimal set of behaviors from a range of potential inputs.
In terms of NPC logic, it should be fairly easy to take as inputs both
internal states (health, mood, etc.) and external 'sensations' from the
NPC's environment to generate the appropriate new internal state, external
action, and/or speech string. One interesting thing about this is you can
have an NPC whose mood, for example, swings slowly from frustration to
anger to fear, generating appropriate (perhaps assembled at run-time)
conversational strings. You can also balance internal states like hunger,
damage, and fear against each other without having to resort to a
table-driven or rule-based approach. The NPC could also learn to add new
behaviors based on clusters of similar experiences.
Mike Sellers Chief Alchemist
mike at online-alchemy.com Online Alchemy
Combining art & science to create new worlds.
More information about the mud-dev-archive
mailing list