A new entropy based approach for robust optimization (application for passenger safety)

摘要

In an attempt to reduce the computational effort involved in a typical robustness design optimisation process, the concepts of "entropy" and "complexity" have been introduced as alternative new measures of system uncertainty. A new optimization problem is defined which simultaneously minimizes the criteria while ensuring an adequate robustness of the optimal solution. This has reduced the computational effort by a factor of 100-1000, depending on the problems investigated. These new concepts have the benefit to optimise designs in a much more coherent, global and efficient manner, permitting the minimization of objective functions whilst delivering optimal system variables and robust system parameters.