Multi-criteria Multi-scenario Approaches in the Design of Vehicles

1. Abstract This paper summarizes findings related to dealing with the complexity of the design of vehicles for optimal dynamic performance under various driving scenarios. It describes how the authors and their students have been able to improve various algorithms and approaches over several years and where they stand today. The work reviewed starts with a MonteCarlo-based vehicle design methodology for obtaining desired vehicle handling characteristics using an eight-degree of freedom nonlinear car model to simulate the vehicle’s dynamic behavior under various driving scenarios: J-turn, Sinusoidal and Double Lane Change maneuvers. Genetic algorithms and simulated annealing are then applied to the same or a similar problem, and their performance compared to the Monte Carlo approach. In parallel to these non deterministic approaches, a mathematical approach based on generating Pareto or other optimal sets of multiple sub-problems and a way to combine them to obtain solutions that are optimal to the combination of objectives and scenarios is presented. The paper concludes with a discussion of the various methods, and their domain of applicability especially when considering the design of vehicles for dynamic behavior.