Development of robot soccer strategies by genetic programming util-ising a simulated environment

Motivation Robot soccer captures the fickle attention of the masses of the great unwashed long enough for money to be filched from their pockets. That is to say, the public finds great interest in the idea of robots playing soccer. In fact, humanoid robots or robots with human-like intelligence generally obtain much attention; the star of the movie Bicentennial Man, or the robotic companions in the cartoon The Jetsons, or 'Data' in the Star Trek series are all examples of this attraction. But the added interest of soccer (especially in Europe) makes this topic unbelievably exciting. At least, it is that excitement which the organisers of the robot soccer competitions hope to generate. The two main competitions (very independent, but similar) are RoboCup and MiroSot. Their current objectives have lower expectations than to see a team of humanoid robots running around after a soccer ball leading 2-1 over the humans after halftime. Nonetheless, [Kitano et al, 1998] makes it clear this is their final, albeit distant, goal. The authors also compare RoboCup with chess—a game 'conquered' by programmers when Deep Blue beat world-champion Kasparov in 1997. While the environment for chess is static, RoboCup is dynamic; Chess is turn-taking while RoboCup is real-time; in soccer, one player cannot see the entire field as a chess player can see their board and finally control is central in chess, whereas RoboCup demands a, more difficult, team effort. The soccer-playing domain is significantly more complex than that of chess and as such makes an excellent challenge. It can be said that much was promised, but very little delivered, in the early stages (circa 1970) of Artificial Intelligence (AI). This lack significantly reduced funding, and hence the work, in the field of AI. The Tokyo workshop in 1992 on Grand Challenges in Artificial Intelligence looked not only for interesting academic work to restore AI's status, but also work that could produce intermediate results that would be of interest outside academia. Perhaps the grand challenge of how to produce funding was solved when the workshop led to the development of RoboCup! Excluding funding issues, the realisation of humanoid soccer-playing robots has some interesting effects. The applicable techniques, to create such a marvel, are anything but obvious and so must be developed. The development will be done in stages—much like a river is not crossed in one big jump, but rather by hopping from stone to …