The weekend roundups at AiGameDev.com are almost back to normal! There’s still a lot going on in the world of game AI, judging from the articles below. The big news on the site this week is the re-opening of the Members’ Area and its continuous training program (finally!); you have slightly over two days left to sign-up. Be sure to swing by The Game AI Forums for some stimulating discussion, and also don’t forget our Twitter account for random thoughts…
This roundup was written by Andrew Armstrong and Alex Champandard. If you have any news or tips for next week, be sure to email them in to editors at AiGameDev.com. Remember there’s a mini-blog over at news.AiGameDev.com (RSS) with game AI news from the web as it happens.
Automatic Game Rule Creation
Julian Togelius sent in a blog post about an interesting presentation he gave at at the recent CIG conference, titled “An Experiment in Automatic Game Design.”
“What we’re trying to do is search a space of game rules for rule sets that constitute fun games. This immediately raises two questions: how do you define and search a space of game rules, and how can you measure whether a game is fun?”
Aleks Krotoski at The Guardian Gamesblog also picks up on the story with her own commentary.
“This week, our interest is piqued by reports of a game developed without human intervention. Julian Togelius, a postdoc researcher in Artificial Intelligence at the Dalle Molle Institute for Artificial Intelligence in Switzerland, has aimed to create a machine that generates a game based on ‘meta-rules’, with the aim of creating something - automatically - that is, well, fun.”
Finally, Slashdot has a discussion on the piece, with some comments.
We’ve got information on a upcoming GDC presentation on multithreaded AI from Intel’s blog:
“I am going to give a presentation about multithreaded AI at GDC this year. We will examine how AI can be threaded and live in a highly parallel environment. How can you thread AI? How can AI talk to physics running on another thread or device? Is deferred processing worthwhile? Do you have to thread your designers? My presentation will hopefully answer these questions and more. The presentation will end with a quick overview of Intel’s Smoke demo; Smoke is a n-way threaded framework that includes source code for highly parallel AI.”
Also, AIIDE-09 calling for speakers and the previously noted AAAI conference proceedings are online if you missed the previous news.
MIT Lectures: Shortest Path Algorithm
A set of three videoed lectures with details, all to do with shortest path algorithms. The first one is above, and covering a wide range of typical solutions:
This is the twelfth post in an article series about MIT’s lecture course “Introduction to Algorithms.” In this post I will review a trilogy of lectures on graph and shortest path algorithms. They are lectures seventeen, eighteen and nineteen. They’ll cover Dijkstra’s Algorithm, Breadth-First Search Algorithm and Bellman-Ford Algorithm for finding single-source shortest paths as well as Floyd-Warshall Algorithm and Johnson’s Algorithm for finding all-pairs shortest paths.
An Introduction to Computer Go
Another learning resource, if you’re interested in Go as a technical AI problem, this is a good starter for learning about the field of using AI to play Go.
The Creator of Creatures Hints His New Work
Creatures creator Steve Grand is noted as being back with this article on his work on biologically based AI called “Grandroids”.
“You may not have heard from Steve lately, but he has been busy—both in artificial life and robotics. Sim-biosis, his underwater life-form simulation is well underway, while his Graindroids project involves building a series of intelligent robots for rent, as crowd pullers in public events and trade shows. His first robot is a five foot tall humanoid female called Grace.”
Autodesk reveals their plans for Kynogon that they brought previously…
“What we’re focusing on is a complete solution for believable characters and that will run from art packages to runtimes,” enthused Michel Kripalani, Autodesk’s senior games industry manager. “With the next generation of gaming platforms, the focus is going to be on runtime simulations. Of course, art tools are also becoming more tied into physics and AI, and while we already had the HumanIK technology, we wanted a complete team that could develop, market, sell and support middleware.”
…morpheme 2.0 has also been released, which boasts NVIDIA PhysX integration…
““With morpheme 2.0, we have developed a method to give programmers and animators much more targeted and differentiated control over physics and animation. It is now possible to add arbitrary physics modes to different parts of the same body – all graphically.”
…and finally Alice McGee’s Grimm is noted to use AI Implant to render behaviors precisely.
“AI implant proved to be essential to getting the AI of our NPCs right. All the NPCs in Grimm have very specific behaviors, and using AI Implant enabled Marwin, our Technical Level Designer, to make those NPCs behave the way they should. He could do this without having to bug the programming team to code new behaviors every time we came up with new requests for NPC AI. This greatly improved both speed and efficiency, and allowed the programmers to focus on all the other tasks at hand.”
AI on your GPU
NVIDIA and ATI went on a PR drive to associate AI development with the GPGPU.
“Nvidia’s director of product management for PhysX, Nadeem Mohammad, agrees,telling Custom PC that ‘all the simple, complex operations’ involved with path finding and collision detection ‘are all very repetitive, so path finding is one of the algorithms which does work very well on CUDA.’ ‘You can always imagine CUDA as loads of processors running the same program but not the same instruction, and ideally on the same data set but with different input parameters,’ says Mohammad. ‘So, in the context of AI, the data set consists of the whole game world, and the parameters going into it are the individual bots – that’s one way of neatly parallelising the problem. If you look at it in that context then any AI program could be accelerated.’
The idea of GPGPU-accelerated AI is also appealing to game developers. ‘I think there’s a lot of potential for GPU acceleration to benefit AI,’ Relic’s senior programmer on Dawn of War II, Chris Jurney, told Custom PC, ‘all our AI is grid based, and we’re already using rasterisation to keep our maps up to date, and for line-draws on those maps to test for passability, so it’s a great match.’
Kokatu raises the point that there is no standardisation on this with ATI, so might be less useful then first expected for most developers.
“Only problem we can see - and it’s a big one - is whether ATI and Nvidia would bother to actually standardise this, or whether we’d end up with two competing solutions that would split the developer community and make the whole thing a royal pain in the ass.”
Videogames as Emotional Simulators?
Earnest Adams has an interesting, if somewhat shallow, article on his thoughts on if games can be realistic emotional simulators (taking examples as Facade above into account).
Far Cry 2, NPC reactions
Gamasutra’s Far Cry 2: Looking Back, Looking Forwards article has some details on the design decisions around NPC reactions.
Question: Though obviously you stressed they’re totally different games, Fallout 3 and Fable II both also have systems that modify NPC reaction to the player based on in-game actions. But In those games, they explicitly set a plus or minus score that occurs at the time you perform an action. In yours, the player isn’t really aware of it unless they look for it.
Answer: […] But it’s not clearer that any one of those measures would have significantly improved it. I think that maybe it’s the wrong issue. For us, it would have been asking the wrong question, because instead what we should be trying to do is seek out those values, those metrics, in the game that are easy for a player to parse just by using their senses. Just by hearing, you’ll see the way the AI is behaving — just by hearing the dialogue, and just by seeing the way the game world alters itself.
AI NodeGraphs in Half Life 2 Developers Commentary (0:50)
For everyone who’s not found it in Half-Life 2, here is the video containing the directors commentary on AI NodeGraphs in Half Life 2.
Behavioral Mathematics for Game AI
Dave Mark’s book Behavioral Mathematics for Game AI is available to pre-order, check out the blurb below to see if it looks to be your fancy for helping improve your AI skills.
“Perfect for intermediate to advanced game programmers, this book shows readers how to use AI programming tools and techniques to create more realistic and interesting behaviors in video games. Readers are shown various ways of using statistics, formulas, and algorithms to create believable simulations and to model behavior. Additionally, the book introduces a number of tools that can be used in conjunction with standard AI algorithms to make it easier to use the mathematical models.”
Godfather 2’s AI Families
Finally, a quote to end the roundup - John Calhoun promises AI run rival families with personalities like real people, we’ll see if this game crops up more with proven AI when it’s released.
“Another cool technical innovation is the AI that runs the strategy game. While you’re playing the game, so are the five rival families you’re up against. Their actions are driven by a system that mimics playing against real people. (In fact, the rule set is derived from a table-top card game we developed!)
These families have personalities - some are vindictive, some are defensive, and some are downright devilish. When they come after you, you’ll realize that their decisions aren’t random. They’re pretty smart, and they know how to hit you where it hurts the most!”