Multi-objective mixed-integer design optimization of planar inductors using surrogate modeling techniques

In this paper, we discuss multi-objective design optimization of planar inductors using surrogate modeling techniques. The goal is to identify the best possible trade-offs between the quality factor of the inductor and its size while maintaining a required value of the inductance at a given operating frequency. The design problem is formulated as a mixed-integer task involving geometry parameters as well as the number of inductor windings. The initial Pareto front is found by optimizing a data-driven surrogate of the structure at hand, further refined by means of response correction techniques. Our considerations are illustrated using a 3.5-nH spiral inductor implemented in 65-nm CMOS technology.