arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2505.03352physics.optics
收藏

无源硅光子学的加速逆向设计

Accelerated inverse design of passive Si Photonics

Anton Sofronov, Dina Yakovleva, Alexander Yakubovich

AI总结:

本文提出基于GPU可微参数化与广义卷积的硅光子无源器件逆向设计方法,结合开源FDTD和梯度下降优化器,将设计时间从数天缩短至分钟/小时级。

摘要:

我们提出了一种精确的可微参数化方案,用于集成硅光子学无源器件的拓扑优化和逆向设计。我们表明,从控制变量到在Yee网格上离散化的有效各向异性介电张量的变换链中的大多数操作,都可以用广义卷积来表示,并且可以在GPU加速器上高效计算。结合最近的开源GPU FDTD代码和具有最终设计可制造性动态控制的自定义梯度下降优化器,该方案形成了一个快速的逆向设计工具,能够在几分钟/几小时内(而不是几天)生成可靠的硅光子器件布局。最终设计虽然并非全局最优,但通常具有物理意义,因为它们反映了场本身倾向于以最高效方式在器件中传播的特性。

英文摘要:

We present an accurate differentiable parametrization scheme for topological optimization and inverse design of passive components of integrated Si photonics. We show that the most of operations in the transformation chain from control variables to effective anisotropic dielectric tensor discretized on the Yee grid are expressible in terms of generalized convolutions that can be computed efficiently on GPU accelerators. Combined with the recent open-source GPU FDTD code and custom gradient-descent optimizer with dynamic control of final design manufacturability, it results in fast inverse design tool capable of producing reliable Si photonic component layouts in minutes/hours instead of days. The final designs, although not globally optimal, are generally physically meaningful since they reflect how fields themselves tend to propagate through the device in the most efficient way.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑