Optimization-assisted Box-Behnken design

Optimization-assisted Box-Behnken design (BBD) combines the Box-Behnken three-level experimental design with a formal optimization step to locate factor settings that maximize, minimize, or hit a target for one or more responses. BBD fits a second-order response surface model using fewer runs than a full factorial, and the optimization stage — typically via desirability functions or numerical search — then exploits that fitted model to identify the true optimum within the experimental region.