[General Algorithme] Working with AI

Got questions? Got answers? Go here for both.

Moderator: MaxCoderz Staff

User avatar
VahnRPG
Extreme Poster
Posts: 331
Joined: Mon 20 Dec, 2004 10:44 pm
Location: Earth...currently...
Contact:

Post by VahnRPG »

well, for that kind of game, I just had it

Code: Select all

ld b,3
call ionRandom
to have it pick which choice the computer made...in fact, for the most part, that's been the way it's been with all my games, just have the computer randomly pick something...only trouble then is that the computer's random generator isn't always random.
Actually, i'm wanting to try to make a more involved computer for Blue World - Quest...but HOW without bloating down the code?
What i WAS thinking about doing was make a pattern list for each monster, then it randomly determines how many of that pattern to do, and from what starting position in the enemy's list...hm...interesting...
I'd rather be a nobody remembered by someone than a somebody remembered by noone.

'Boku' can't save the world - Blaze Rennaul
--'Boku' doesn't have to... - Klyne Ryuuno
Blue World - Coming soon
User avatar
Matthias1992
New Member
Posts: 16
Joined: Mon 13 Jul, 2009 6:23 pm

Re: [General Algorithme] Working with AI

Post by Matthias1992 »

Perhaps combining the power of SMC and the speed of ASM coudl produce notable "Intelligence". my most recent thought about A.I. Was to base the intelligence not on preprogrammed (and thus limited) awareness but on chance, what? chance? yea, you heard me right, chance. As has been made obvious in the previous posts there are limited situations in which an NPC might come. Now depending on data from the player (position, sound, stealhtyness etc) you could calculate chance indexes for one of these events (or others) to happen. Now comes the nice part, instead of acting in manner A when event B occurs the A.I. will try to prevent situation A from occuring by enabeling manner C. i hope it is not all too confusing but globally said: Notable A.I. can be made (according to me :P) by making chance indexes, now the A.I. will react before an event actually happens, based on the chances of it to happen. This makes A.I more layered then regular "If this happens then do that" A.I. The SMC part of this might be in the manner It reacts to the same probability of an event to occur. Now, all of this is sheerly theorethical of course :P
Post Reply