Jump to content
  • Advertisement


  • Content Count

  • Joined

  • Last visited

  • Days Won


IADaveMark last won the day on July 11

IADaveMark had the most liked content!

Community Reputation

3817 Excellent

About IADaveMark

  • Rank
    Moderator - Artificial Intelligence

Personal Information


  • Twitter
  • Twitch
  • Steam

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

  1. Not really for standard behavioral AI. The problem is primarily that it doesn't scale well as you add more potential actions with objects. The search space explodes geometrically as you add more things to do. Also -- and likely worse -- you have to re-plan anytime anything in the state space changes and invalidates one of your nodes. So not only do you have huge search times, most of that is going to get thrown away on many of your think cycles. Now multiply that across a lot of agents.
  2. 1024 nailed it. People were fooled into thinking the AI was better than it was. The stuff that happened in there that is GOAP based is easily duplicatable in plenty of architectures such at BTs or utility.
  3. IADaveMark

    Unity Asset -Visual FSM Editor

    Most people can write an FSM implementation (non-graphical) in about 2 hours. Graphical ones are a bit of overkill since all they are doing is showing the same transition logic between states X and Y that you are putting in your code of state X -- often a line or 2. Therefore this does seem a little much. *shrug*
  4. Define "advanced". And for that matter, define "AI".
  5. IADaveMark

    AI for Top Down Kinda-Minimalist RTS?

    AI is often the most complicated programming of a game. And RTS AI is most often the hardest type of AI. You have your work cut out for you. Figure 10 or 12 things to learn before you even start writing real RTS AI.
  6. IADaveMark

    Little Help for an AI/Bot for a game

    Um... no. If he can't even do image processing to determine what the map is, influence maps aren't going to help him at all. (And I've written articles and GDC lectures on influence maps and use them in my consulting work on AAA games... so I've kind of been there.)
  7. IADaveMark

    Little Help for an AI/Bot for a game

    If you aren't really into this "AI stuff" then this project is far too complex for you to start with... and not just because of the AI part. The harder part here is doing things like image recognition to recognize the map and the units, understand the board, process the spatial aspects, etc. And, from the sounds of it, you don't even know how to program and execute a simple AI decision even if you had all that information already. That last part should be your starting spot... not the ending one. Best bet is to learn how to do simple things like doing the AI for Tic-tac-toe or other simple board games where the information about the game state is already right there in front of you in game data. Once you have done that a few times, then getting into the advanced stuff of doing image recognition, etc. you might be able to tackle.
  8. To be fair, an "AI Designer" is more designer than programmer. However, a lot of designers can't handle thinking in the manner that is necessary for crafting AI behaviors. Think of it as a technical specialty like a "technical artist" is. Yes, an artist but with a particularly technical bent.
  9. I know you asked for "technical skills", but TBH, one of the most under-recognized skills for an AI programmer is thinking about behavior rather than code. That said, there is serious room for an AI programmer vs. an AI designer. In fact, I would suggest that there is a difference between and AI architect and an AI programmer. You don't have to be a killer programmer to be an AI architect... just know how it all fits together on the higher level to serve the needs. Similarly, an AI designer tends to work more with the behavior data and crafting how it is working on-screen. It certainly helps to have programming skills in both of those, but they pair well with a true "AI programmer".
  10. We need to clear up some terminology here that seems to be throwing some people off. The main difference between a state machine (i.e. FSM) and other things is, as mentioned, the transition logic out of a state and to another one is inside each individual state. The main problem that creates is that, when you create a new state, all logic for getting to that new state needs to be added to all of the existing states that could move to the new one. So if you have 20 states and you add a new one that is so high priority that it could jump out of all 20 of the other ones, you now have to put the transition logic in 20 different times/places. In the worst case scenario, if every state could lead to every other one, you get a massive n2 problem. It's a bitch to maintain. What behavior trees (and subsequently other architectures) did was remove the transition logic from the states and put it into a stand-alone reasoner. That is, evaluate all of the issues in one place and then decide what state to be in. Note... all of these other architectures still end up in a "state". That doesn't mean they are "state machines", however. In fact, Jeff Orkin's paper on GOAP was titled, "3 states and a plan" because his states were quite literally simply moving somewhere, doing an animation, and using a smart object. My IAUS is similar in that it's "state" is whatever it is doing at the time. However, like a BT, the reasoner is separate.
  11. For (repeated and continual) reference. http://intrinsicalgorithm.com/IAonAI/2012/11/ai-architectures-a-culinary-guide-gdmag-article/
  12. IADaveMark

    Context Steering

    I believe that you may be looking for something that is considered so simple that no one has bothered to write code for sharing. Really everything that he is talking about is conceptual variations onto existing technology. While it is certainly a novel idea, it really wouldn't be hard to implement. What is your level of experience?
  13. IADaveMark

    Context Steering

    Did you watch this? That's what Apoch refers to in his article. https://gdcvault.com/play/1018262/The-Next-Vector-Improvements-in There's probably not code in it (I don't remember despite running that Summit for 10 years) but it might help wrap your head around how to write your own code.
  14. IADaveMark

    Context Steering

    This used to be the go-to site for that. See if the java code is still there. https://www.red3d.com/cwr/steer/ Looks like the Java stuff is gone (as are his great demo vids), but this should help. http://opensteer.sourceforge.net/
  15. Well, considering that I am the guy that, for all intents and purposes, spearheaded and popularized utility AI (along with Kevin Dill), and there are game studios around the world trying to implement my IAUS based only on my GDC AI Summit lecture about it (with @ApochPiQ), I imagine it would be well received as THE utility system for Unity (and anything else, for that matter). And no, I don't build it using the editor. It's entirely data-driven externally. Now that I'm finally getting over almost dying at GDC 14 months ago, maybe I can get some things in order.
  • Advertisement

Important Information

By using GameDev.net, you agree to our community Guidelines, Terms of Use, and Privacy Policy.

GameDev.net is your game development community. Create an account for your GameDev Portfolio and participate in the largest developer community in the games industry.

Sign me up!