A Genetic Algorithm Application in Railway Engineering

摘要

The aim of the paper entails showing the methodology developed by the authors to obtain wheel profile geometries adapted to running conditions of railway vehicles, in order to get better dynamic behavior of the vehicle when running over the tracks. In this way, an improvement of riding performances with reference to some physical phenomena (as wear, wheel-rail contact stresses and risk of derailment) will be achieved. This methodology uses the Genetic Algorithms technique, which is based in Darwin's evolution theory, in such a way that the individuals that are better adapted to its environment are the ones which get descendants. This process leads to an improvement of the species in the real world, and it should lead to an improvement of the wheel profile performance for this application. In this kind of technique, populations of individuals are considered (in this case each individual is a wheel profile geometry'). In the application case presented, the adaptation to the environment is measured by means of indexes that are related to dynamic behavior of the vehicle running over the track, when it is provided with a specific wheel profile.