Use of Genetic Algorithms as an Innovative Tool for Race Car Design

Design processes of modern race car are often developed in short time; during this period a large number of parameters has to be tuned to reach best results. Many kind of vehicle dynamics simulation models have been developed by car manufacturers and private suppliers to investigate car behavior on racing tracks. Such models have a high degree of complexity but they can be employed, in a simplified mode, during design processes of racing cars for which tracks technical data (for such circuits where they will race) are well known. During the first stages of the design activity very complex numerical models (such as multibody simulations) are not necessary and it is possible to use simplified methods to locate optimal solutions in a fast way. In present work a numerical model, able to reproduce car behavior on a defined circuit or simply analyze meaningful test cases (breaking, steering,…), is used to appraise performances of race car with different technical configuration. Besides is proposed a method to reduce designing field of investigation through genetic algorithms (GA's) using vehicle numerical model to individuate such solution that gives best performances on each circuit. Numerical algorithm is tested to choose best from a large number of technical configurations on different tracks. Several types of genetic algorithms are used: results show that GA's can be a useful tool to fasten design processes narrowing the field of investigation. In some cases the optimization method developed gives analogous results as what would be carried out by an experienced engineer.