Figure 2
From: Level set methods for gradient-free optimization of metasurface arrays
A flowchart for a general topology optimization problem. (a) A smooth material distribution function \(\rho\) is defined on a subset \(\Omega\) of \(\mathbb {R}^2\). The contour lines represent regions of constant material. For binary designs, the material can be thresholded based on the zero level (red). (b) The forward operator \(\mathscr {H}\) is problem dependent but typically consists of a PDE/ODE simulation with pre- and post-processing. (c) The simulated system response \(\mathscr {H}(\rho )\) is compared against a desired system response \(\varvec{y}\). (d) The simulated and desired responses are fed through an objective function to provide feedback to an optimizer. The optimizer then updates the material distribution function for the next evaluation.
