Game AI Roundup Week #9 2009: 9 Stories, 1 Job, 1 Video

Andrew Armstrong on March 2, 2009

Here we are with another Game AI Roundup here at Be sure to swing by The Game AI Forums for some stimulating discussion, and also don’t forget Alex’s Twitter account his random thoughts…

This roundup was written by Andrew Armstrong (site). If you have any news or tips for next week, be sure to email them in to <editors at>. Remember there’s a mini-blog over at (RSS) with game AI news from the web as it happens.

Artificial Intelligence in The Sims 3

Sims 3 AI Graveyard

Richard Evans from EA brings some information about how the AI in the Sims 3 is looking like. Not many technical details, but it is a title to look out for when investigating life simulation AI when it eventually comes out.

“Hello everyone! My name is Richard. I’m the AI lead on Sims 3, and I’ve been busy working on the Sims’ free-will – what they decide to do when you leave them to their own devices.

One big thing we have insisted on from the start is that the Sims won’t need you to hold their hand every time they need to use the toilet. They are now able to take care of their basic needs on their own, leaving you to focus on more important things: building relationships, expanding your career opportunities, and exploring the world around you.

But perhaps more importantly than solving the base physical needs, our major focus has been getting the Sims to express their individuality. Each Sim has a personality defined from a large array of traits, and these traits have a strong influence on autonomous behavior. A flirty Sim will preen and pose, while an un-flirty Sim will reject all but the most persistent suitor. Some Sims are family-oriented – you will see them playing games with their children and cooing over them; others dislike children, and can be heard complaining about them loudly. All the Sims will express their personalities on their own, without needing constant input from you.”

AIIDE Call for Papers (Final)


Kevin Dill, resident expert and AIIDE Organizational Committee Member, sent in details about the call for papers this year. The deadline is April 14th.

Research Track
Research Track papers describe AI research results that make advances towards solving known game AI problems or enabling a new form of interactive digital entertainment. The novel technique should be validated in a game prototype or test-bed, but need not be validated in a commercial game. Research Track papers are evaluated by the highest standards of academic rigor. The highest rated papers will be presented in short lecture format. The next highest rated group of papers has the opportunity to present their work in a poster session. Applicants submit a paper of no more than 6 pages in the AAAI format for blind review (i.e. authors names and affiliations are omitted). All papers will be allocated 6 pages in the proceedings regardless of presentation format.
Industry Track
Individuals that have game development experience but lack the time or need for publishing rigorous academic papers can alternatively apply to the Industry Track. This track will include presentations of AI techniques, issues, or case studies from the perspective of implementing a product in the current commercial environment. Presentation proposals will be evaluated on their potential for conveying clearly elaborated ideas that have not been previously described to an adequate degree. Industry Track applicants submit an extended abstract describing the content of the proposed talk that also includes one paragraph describing their game industry experience. An extended abstract of two pages is sufficient, although any length up to that of a full paper (6 pages) is acceptable. Abstracts will be published in the conference proceedings.

New AI details for EVE Online

EVE Online

From this developer blog post, we find out more details of the changes to the AI in the MMO. It appears that the bots will play more in the style of players, which will likely be a nice change of pace versus the original very basic FSM AI.

“One of goals we have worked towards is making PvE combat more like PvP combat. This has a few benefits. First, new players get a taste of what PvP is like while still doing PvE encounters. Second, they will have ship fittings that are more suited towards PvP, which means that when the victi… um, player gets jumped by another player he has a fitting that will allow him to fight back effectively, or at least run away. Lastly, it will make for more engaging combat that is both dynamic and feels alive.”

AI Programmer Job at Kaos Studios

Kaos Studios

A job at Kaos Studios this week for an AI Programmer.

Kaos Studios is located in the heart of New York City and is mere blocks from the Empire State Building and the thrill of Midtown Manhattan. Along with the opportunity to live in one of the most exciting cities in the world, we also just finished up one of the most exciting FPS titles to date. Frontlines: Fuel of War (PC/XBOX360) is already receiving great press and that’s just the beginning! We also offer competitive salaries, comprehensive health benefits, and an excellent compensation package. We are always looking for talented artists, developers, and designers to join our growing team, so check out our job postings and let us know what interests you!

On Resident Evil 5’s AI

This is a producer, but still, it is interesting to see what changes Resident Evil 5 has had in it’s design related to the AI. Zombie AI is going to be one hot topic at GDC at least!

“BF: One horror game I played recently that very much impressed me and a lot of people I know is “Left 4 Dead.” The reason being that the artificial intelligence is the main reason that game was scary because you never knew what was going to happen. If you played that game, what did you think of it and the A.I.? And relatedly, are there any significant changes in the A.I. of the enemies in “Resident Evil 5?”

JT: Yes certainly I have played that and a lot of people at Capcom are playing it. It’s a great game, a lot of fun, it’s really well put together. As you said, the A.I. system is very interesting. I think that element in creating that kind of game is really important because you don’t know what’s going to happen, where the enemies are going to come from. I think they did a great job creating a shooting game with zombies that’s very different from “Resident Evil” or any other zombie related games out there. One of the things I think is really interesting about “Left 4 Dead” is that system works because the zombies are always running at you, always chasing after you. It works really well in that game. As regards the A.I. system in “Resident Evil 5,” it has evolved and improved over “Resident Evil 4.” Probably the biggest change would be that there are lots of different environments in the game and no matter what environment you put the A.I. into, they don’t end up getting lost or not being able to find you. They’re always able to come after you and attack you. That has been a big improvement.”

Designing the Framework of a Parallel Game Engine

Parallel Game Engine Design

Jeff Andrews writes a sponsored Gamasutra article on Parallel Game Engine creation, detailing one reason for doing so might be with AI:

The lock step mode can also implement a pseudo-free-step mode of operation by staggering calculations across multiple steps. One use for this might be with an AI that calculates its initial “large view” goal in the first clock, then instead of just repeating the goal calculation for the next clock comes up with a more focused goal based on the initial goal.

Bot Navigation With No Items To Search For

Ted Vessenes posts up some thoughts on navigation AI for bots, in Quake 3-like games when there are no items in a level.

“Where do you go when there are no items on the level?

This might not seem like a big deal at first glance. But remember that bots use items as the primary motivating factors for deciding where to go. The pathing and routing code will tell you how to get from point A to point B, but all of that is meaningless if you don’t know what point B is. The typical strategy is for the bot to pick up the items that help it the most and require the least amount of movement. If the bot can’t find any enemies, it will head to the nearest generally useful item and hope a target wanders by. Item placement is the core component of goal selection in BrainWorks.

If you take away the items, everything falls apart.”

Pac-Man’s AI

Pacman Scatter

A deeply insightful investigation of every single behavior coded into Namco’s massively successful Pac-Man - the game is an early example of rather more subtle and complex AI then you would have thought, where difficulty is increased by improved AI algorithms not just throwing more enemies out :)

“What design and AI lessons can we learn from Namco’s seminal Pac-Man? From history through behavior, Gamasutra presents a comprehensive Jamey Pittman-authored guide to the classic game.”

In The World of Software Architecture…

The regular social media channels coughed up the following articles this week.

Wavefront Algorithm for Robots

Many algorithms applied in game AI today are historically from robotics research. This interesting tutorial explains the wavefont algorithm.

“The theories behind robot maze navigation is immense - so much that it would take several books just to cover the basics! So to keep it simple this tutorial will teach you one of the most basic but still powerful methods of intelligent robot navigation.”

Oh, You Newspaper Lot…hehehe!

Lastly, the inevitably poor reporting from The Sun (If you’re not from the UK, it’s kinda akin to a children’s book but with naughtier pictures ;) ) brings this gem of a quote related to videogame AI, relating it to their feature on the Toshiba robot. Can anyone think of any better examples? You know, ones not over 10 years old? :)

“The artificial intelligence in MicroProse’s Civilization II is generally reckoned a great success. Other games have not been so fortunate.”

Discussion 0 Comments

If you'd like to add a comment or question on this page, simply log-in to the site. You can create an account from the sign-up page if necessary... It takes less than a minute!