Multiobjective optimization design for an occupant restraint system considering interval correlation

The occupant restraint system (ORS) focuses on automobile crash safety, which can effectively reduce passenger injury. For an ORS, some design parameters are uncertainties, but even they may have correlations. It is noted that the ORS is typically characterized by various occupant injury indices, and uncertainties and their correlations will impact these indices, while most existing design problems have been formulated as a single objective optimization. To address these issues, this paper combines a multidimensional parallelepiped model with a multiobjective genetic algorithm (GA) for the vehicle ORS design. First, a multiobjective optimization model for the ORS design considering interval uncertainties and their correlations is developed to balance the design requirements for multiple objectives. Second, the established multiobjective optimization model considering parametric correlations is converted into an independent interval multiobjective optimization model in the transformed cuboid domain. Third, the interval multiobjective optimization model is converted into a deterministic multiobjective optimization model by the use of the interval order relation and interval possibility degree. Finally, the optimization problem is solved by coupling the multiobjective GA with the interval expansion method. The ORS design of a 100% frontal impact at a speed of 35 mph is exemplified for the proposed model and method, and the application example shows that the optimization results are different for different correlations. If the correlation analysis is neglected, then the optimization results may lead to a deviation of design. Thus, more conservative solutions could have been generated from the ORS design considering interval correlation.

Automated summary

This paper combines a multidimensional parallelepiped model with a multiobjective genetic algorithm (GA) for the design of the vehicle occupant restraint system (ORS). By considering interval uncertainties and correlations, a multiobjective optimization model is developed, and the optimization problem is solved using the interval expansion method. The application example shows that correlations have an impact on the optimization results, and neglecting correlation analysis may lead to design deviations.