ABSTRACT Multilayer thin films are fundamental components of photonic and optoelectronic technologies, yet their inverse design and characterization remain limited by the trade‐off between exploration of the solution space and computational cost. This paper proposes a tensorized quantum genetic algorithm (tQGA) with a selective evolution strategy, in which each individual evolves independently toward probabilistically chosen targets, maintaining diversity while ensuring stable convergence, and thereby enhancing optimization performance. A tensorized implementation further enables parallel updates of the population and simultaneous optical calculations for all solutions within each generation, achieving a 60–90× speedup over conventional frameworks, and up to 300–500× with GPU acceleration. The proposed tQGA is validated across three representative thin‐film design and characterization tasks, consistently demonstrating superior accuracy, robustness, and computational efficiency. These results clearly demonstrate the significant potential of tQGA as a general and efficient framework for addressing inverse problems in thin‐film optics.
Liu et al. (Sat,) studied this question.