Optimization Algorithms for tuning suspension systems used in ground vehicles
Metadata only
Date
2002Type
- Conference Paper
ETH Bibliography
no
Altmetrics
Abstract
According to bibliography many researchers have contributed in the area of optimization of the suspension characteristics used in ground vehicles. In a number of papers different versions of Evolution Strategies are implemented for achieving this goal. A disadvantage of such methods is the large number of iterations for the convergence of the algorithm. We propose the acceleration of evolution strategy with means of deterministic algorithms, such as the steepest descent method, and we prove that such combination yields significantly faster and more reliable convergence. The method combines the advantages of two categories of optimization algorithms, deterministic and stochastic. The half-car model of suspension systems, described with non-linear differential equations and subject to various road profiles, constitutes a suitable basis for the implementation of numerical optimization algorithms. The common objective of the presented optimization procedures is the improvement of the passengers' ride comfort, leading to minimization of the maximum acceleration of the sprung mass, with respect to the geometrical constraints of the suspension as well as the necessary traction of the vehicle. The comparison between the performances of the implemented methodologies states clearly the advantages of the proposed algorithm when the problem is of higher complexity, meaning that the number of design variables is increased and also many several local minima of the objective function are introduced. In addition, plans and topics for further research are reported. Show more
Publication status
publishedBook title
Proceedings of the 2002 SAE International Body Engineering Conference and Automotive & Transportation Technology Conference on CD-ROM (IBAT2002CD)Pages / Article No.
Publisher
SAEEvent
Organisational unit
03890 - Chatzi, Eleni / Chatzi, Eleni
Related publications and datasets
Is identical to: https://doi.org/10.4271/2002-01-2214
More
Show all metadata
ETH Bibliography
no
Altmetrics