SUMMARY The objectives of this research were to investigate genetic algorithms in multi-criteria, multi-scenario optimisation and to implement a feasible optimisation process for vehicle handling behaviour. For this purpose a BMW vehicle model has been coupled with a genetic algorithm (GA) specially designed for optimisation. The design variables of the vehicle model consist of both discrete and continuous variables (for example the mass of the vehicle and the suspension characteristics). More than one hundred and fifty design variables and eighteen dynamic indicators were used in this study. For optimisation, the variation in vehicle design parameters was restricted to a range of 15% and for suspension parameters to 50% around the nominal value. The optimisation has been performed for three different manoeuvres: J-turn, steady state cornering and driving on rough road. All of these manoeuvres were executed using the ISO specifications. Significant improvements were obtained in all the dynamic indicators (up to 89% in some). The results indicate that the use of a GA is a valid approach to multi-criteria, multi-scenario optimisation.
[1]
H.-P. Krüger,et al.
A WORKLOAD APPROACH TO THE EVALUATION OF VEHICLE HANDLING CHARACTERISTICS
,
2000
.
[2]
Giampiero Mastinu,et al.
Optimising a Car Chassis
,
1999
.
[3]
Alexandra Neukum,et al.
A DRIVER MODEL AS DEVELOPMENT TOOL FOR VEHICLE HANDLING DESIGN
,
1999
.
[4]
David E. Goldberg,et al.
Genetic Algorithms in Search Optimization and Machine Learning
,
1988
.
[5]
As.
Structural Optimization: Status and Promise
,
1993
.
[6]
Manohar P. Kamat.
Structural optimization : status and promise
,
1993
.
[7]
Giampiero Mastinu,et al.
On the integrated design of the tyre-suspension system of a racing car
,
2000
.
[8]
Giampiero Mastinu,et al.
OPTIMAL AND ROBUST DESIGN OF A ROAD VEHICLE SUSPENSION SYSTEM
,
1999
.