Road Profile Identification by Genetic Algorithm with Multistage Search using Vibration Response of a Quarter Vehicle Model

Abstract

Vehicle performance is affected by road profile since the road profile supplies vibration to the vehicle. The road profile can be identified by vibration response of the vehicle. By using the vibration response, road profile identification can be, in fact, formulated as a single-objective optimization in which an objective, a minimization criterion, is the numerical difference between vehicle vibration response of actual road profile and that of predicted road profile. This paper present multistage search in genetic algorithm (GA), an optimization algorithm, in the road profile identification. In the multistage search, a solution is divided into a number of parts and each part is consequently evolved as GA process separately from other parts. A quarter vehicle models with two test cases of double bump on road profiles are used. Simulation runs reveal that the multistage search can enhance performance of GA. In addition, the multistage search using the least number of decision variables in each solution part gives the best results of the optimized solutions.