rec.autos.simulators

by **Uwe Schürkam** » Tue, 22 Jun 2004 20:15:00

Hi folks,

interesting article mentioned on /. today:

http://www.racesimcentral.net/,2554,63900,00.html?tw=wn_5techhead

"smack-pot writes "Wired News has an article about how the Digital
Biology Interest Group at University College, London is using genetic
algorithms to breed superfast Formula-One race cars. 68 design
parameters were configurable in the cars, and the generated designs
were tested using the racing simulation software developed by the game
developer Electronic Arts. According to the research it is possible to
shave off 88/100th of a second per lap by using genetic algorithms to
tune the cars. In an industry where a tiny fraction of a second
matters, that's significant."

Looks like they used F1C to test the "genetic" results!

Cheers,

uwe

--
mail replies to Uwe at schuerkamp dot de ( yahoo address is spambox)
Uwe Schuerkamp //////////////////////////// http://www.racesimcentral.net/
Herford, Germany \\\\\\\\\\\\\\\\\\\\\\\\\\ (52.0N/8.5E)
GPG Fingerprint: 2E 13 20 22 9A 3F 63 7F 67 6F E9 B1 A8 36 A4 61

by **Klin** » Tue, 22 Jun 2004 23:42:05

"Uwe Schrkamp" wrote ...

Interesting concept, and thanks for the link, but without real-world testing
perhaps all they have achieved is a set of parameters that works well in
F1CS. Is the physics model within that game really sophisticated enough to
predict actual performance that closely? I wonder if some teams will follow
the same approach in their own testing simulations - or are already?

...Klinn

by **J. Todd Wass** » Wed, 23 Jun 2004 03:51:34

>Date: 6/21/2004 9:42 AM Central Daylight Time

>"Uwe Schrkamp" wrote ...

>> Hi folks,

>> interesting article mentioned on /. today:
>> http://wired.com/news/autotech/0,2554,63900,00.html?tw=wn_5techhead
>>"...is using genetic algorithms to breed superfast Formula-One race cars."

>> Looks like they used F1C to test the "genetic" results!

>Interesting concept, and thanks for the link, but without real-world testing
>perhaps all they have achieved is a set of parameters that works well in
>F1CS. Is the physics model within that game really sophisticated enough to
>predict actual performance that closely? I wonder if some teams will follow
>the same approach in their own testing simulations - or are already?

>...Klinn

This isn't really a new idea. Genetic algorithms are fit for exactly this type
of thing and have been known for many years now. Thought of the same race car
application myself quite awhile ago :-P
Anyway, if a team's F1 simulation is sufficiently accurate then I'd be
surprised to find out some haven't been doing this for some time already. If
not, they're a bit behind the AI game. At the same time, it would be
surprising if the simulations were accurate enough to take advantage of genetic
algos in this application.

Todd Wasson
Racing Software
http://PerformanceSimulations.com
http://performancesimulations.com/scnshot4.htm

rec.autos.simulators

Interesting article on "breeding" race cars

Interesting article on "breeding" race cars

Interesting article on "breeding" race cars

Interesting article on "breeding" race cars