This is the current list of competitions which will be held at CIG 2012. Please contact the respective organizers or the competitions chair if you have any question about them or want to participate.
Kyung-Joong Kim and Sung-Bae Cho has assembled useful resources and tutorials on how to get started with participating in the various competitions on their game AI portal.
The 2K BotPrize competition
Can you make a bot that plays like a human?
Computers are superbly fast and accurate at playing games, but can they be programmed to be more fun to play - to play like you and me? People like to play against opponents who are like themselves - opponents with personality, who can surprise, who sometimes make mistakes, yet don't blindly make the same mistakes over and over.
The BotPrize competition challenges programmers/reasearchers/hobbyists to create a bot for UT2004 (a first-person shooter) that can fool opponents into thinking it is another human player. The competition has been sponsored by 2K Australia since 2008, and the $5000 major prize is yet to be claimed.
For details of how to create your own bot and enter the competition, go to http://www.botprize.org.
*** There will be a warmup competition modelled on the BotPrize at WCCI 2012 in Brisbane, Australia, in June.
Why not try your bot-writing skills there and sharpen up your BotPrize bot? http://www.botprize.org/HumanlikeBots.html ***
organizer: Philip Hingston (sponsored by 2K Games)
Simulated Car Racing Championship
The goal of the championship is to design a controller for a racing car that will compete on a set of unknown tracks first alone (against the clock) and then against other drivers.
The controllers perceive the racing environment through a number of sensors that describe the relevant features of the car surroundings (e.g., the track limits, the position of near-by obstacles), of the car state (the fuel level, the engine RPMs, the current gear, etc.), and the current game state (lap time, number of lap, etc.). The controller can perform the typical driving actions (clutch, changing gear, accelerate, break, steering the wheel, etc.)
Realtime Strategy (RTS) games are one of the major computer game genres and one of the few for which AI-based players (bots) have little chance to win against expert human players - if they are not allowed to cheat. StarCraft (by Blizzard) is one of the most popular RTS games of all time, and is known to be extremely well balanced.
The competition enabled researchers to evaluate their AI systems in this robust commercial RTS environment, concentrated on a tech-limited Terran force 2-player battle.
organizers: Mike Preuss and Tobias Mahlmann
Physical Travelling Salesman Problem
The Travelling Salesman Problem (TSP) is one of the best known optimisation problems in operations research. The Physical Travelling Salesman Problem (PTSP) adds an interesting twist to the TSP: the salesman is now physically embodied and can be moved across a two-dimensional map by applying sideway rotations and throttling.
The goal of this single-player real-time game is to navigate a set of waypoints as quickly as possible, circumventing any obstacles that are in the way. The PTSP thus requires participants to find an optimal path connecting all waypoints while steering the salesman, making it a challenging and non-trivial game where both planning and reactiveness are required.
Ms Pac-Man vs Ghost Team Competition
For the first time, we are happy to introduce a Ms Pac-Man competition where you may submit AI controllers for either or both Ms Pac-Man and the ghost team!.
Ms Pac-Man is a highly enjoyable yet challenging game that is usually viewed form the perspective of the Ms Pac-Man agent, and this has already been the subject of an on-going series of competitions. However, the game also provides an excellent environment for testing multi-agent strategies, by controlling the team of ghosts; here the aim could be either one of optimising the playability of the game, or simply to minimise the score obtained by Ms Pac-Man. It is this latter (more objective) goal that provides the focus of this competition.
Ms Pac-Man Screen Capture Competition
The aim of this competition is to provide the best software controller for the game of Ms Pac-Man. This is a great challenge for computational intelligence, machine learning, and AI in general.
Unlike Pac-Man, Ms. Pac-Man is a non-deterministic game, and rather difficult for most human players. As far as we know, nobody really knows how hard it is to develop an AI player for the game. The world record for a human player (on the original arcade version) currently stands at 921,360. Can anyone develop a software agent to beat that?.
The Ms. Pac-Man competition will test the ability of computer-based players at the conference. We are especially interested in players that use computational intelligence methods to address the problem, but the contest is open to any type of algorithm: you can hand-program it as much as you like.
The mode of interaction is as follows: about 15 times per second your program will be sent a pixel map of the Ms. Pac-Man window, and it then responds with an integer indicating the direction of the joystick.
Mario AI Championship: Gameplay and Learning Tracks
The Gameplay track of the competition is about creating a controller that plays Infinite Mario Bros as well as possible. In the learning track, you must submit a controller that learns to play a particular level as well as possible.
Mario AI Championship: Level Generation Track
The level generation track of the competition is about creating procedural level generators for Infinite Mario Bros.
Competitors submit level generators that generate Mario levels according to certain specifications. We will have a number of testers who will play a test level on-site at the competition event. While playing, the game play behavior of each tester will be recorded. This recorded information serves as an input upon which the competing level generators are going to build their customized levels. The generated levels are then played by the testers, and ranked according to which they liked the best. The winner of the competition is the one who submits the level generator that makes the levels that the judges find the most entertaining.
Mario AI Championship: Turing Test Track
The Turing test track of the competition is about creating Mario controllers that play the game in a human-like manner. The winner is decided by the vote of the audience, who gets to watch videos of human and computer players who play the same levels, and try to tell the humans from the algorithms.
Version 2.5 - September, 2012