Evolving Racetrack Knowledge in a Racing Game.

Abstract

This paper discusses the use of a Genetic Algorithm (GA) to evolve waypoint-based racing lines for computer-controlled cars in an online game. One goal is the reduction of developer effort, in order to automatically produce varied and effective racing lines for numerous different vehicle configurations on various racetracks. Another goal is to assess the design of fitness functions in order to evolve varying strategies and driving styles for the computer-controlled drivers. The effectiveness of this approach is assessed, and the evolved driving strategies on a number of specific racetracks are analysed.