Techniques for Addressing Saddle Points in the Response Surface Methodology (RSM)

Industrial systems have multiple parameters that control the processes and one or more output performance metrics that one would like to maximize or minimize. The design-of-experiment (DoE) techniques have been proposed, which allows the system designers to optimize the control parameters with a reasonable number of experiments, the response surface methodology (RSM) being among the primary of such techniques. In this paper, we use the RSM to optimize the design parameters in a multi-layer microwave-assisted magnetic recording (MAMR) medium. The MAMR is one of the technologies proposed to extend the life of magnetic recording and the multi-layer MAMR (ML-MAMR) or 3D-MAMR, and it has the potential to increase the recording densities further by stacking multiple recording layers on top of each other. The danger of overwriting information in one layer, when targeting the writing of another layer, is of concern; hence, the optimization of the parameters to minimize this is of paramount importance for the successful implementation of the 3D-MAMR. In this paper, we report the techniques tried and used to overcome the saddle points that were encountered during the DoE.