Multi-objective control optimization for semi-active vehicle suspensions

In this paper we demonstrate a method for determining the optimality of control algorithms based on multiple performance objectives. While the approach is applicable to a broad range of dynamic systems, this paper focuses on the control of semi-active vehicle suspensions. The two performance objectives considered are ride quality, as measured by absorbed power, and thermal performance, as measured by power dissipated in the suspension damper. A multi-objective genetic algorithm (MOGA) is used to establish the limits of controller performance. To facilitate convergence, the MOGA is initialized with popular algorithms such as skyhook control, feedback linearization, and sliding mode control. The MOGA creates a Pareto frontier of solutions, providing a benchmark for assessing the performance of other controllers in terms of both objectives. Furthermore, the MOGA provides insight into the remaining achievable gains in performance. ► We demonstrate a method for determining the optimality of control algorithms based on multiple performance objectives. ► A multi-objective genetic algorithm (MOGA) is used to establish the limits of controller performance. ► This method provides insights into controller tuning; it can be used to quantify remaining performance benefits. ► While the approach is broadly applicable, this paper focuses on the control of semi‐active vehicle suspensions. ► The two performance objectives considered are ride quality and dissipated power.