Table 2 Optimization results for Structure I.

From: Low-cost quasi-global optimization of expensive electromagnetic simulation models by inverse surrogates and response features

Optimization method

Algorithm of Section “Globalized search using inverse regression surrogates and response features” (this work)

PSO

Trust-region gradient-based algorithm

20 iterations

50 iterations

Average objective function value (dB)

 − 20.7

 − 14.8

 − 21.9

 − 12.7

Computational cost$

73.3

200

500

63.8

Success rate#

10/10

8/10

8/10

6/10

  1. $The cost calculated as the number of EM simulations of the structure under design, averaged over ten algorithm runs.
  2. #Number of algorithms runs with the operating frequency allocated with sufficient accuracy, i.e., to satisfy U(x*,F) ≤ Ulocal.
Back to article page