We describe an approach to coronagraphic focal-plane wavefront control that utilizes gradient-based nonlinear optimization along with analytical gradients obtained with algorithmic differentiation to find deformable mirror solutions. In addition to eliminating the cost of calculating a high-dimensional finite-difference Jacobian matrix, we show that this approach leads to improved asymptotic computational efficiency. With very high-actuator deformable mirrors such as the 128 × 128 actuators baselined for the Large UV/Optical/IR Surveyor mission concept, the proposed algorithm reduces memory consumption by approximately 95 % compared to a Jacobian-based algorithm.

中文翻译：

---

使用非线性优化的无雅可比冕波波前控制

我们描述了一种冠状焦平面波前控制方法，该方法利用基于梯度的非线性优化以及通过算法微分获得的解析梯度来找到可变形的镜面解。除了消除计算高维有限差分雅可比矩阵的成本外，我们还表明，该方法可提高渐近计算效率。借助非常高的致动器可变形反射镜（例如针对大型UV /光学/ IR Surveyor任务概念设定的128×128致动器），与基于Jacobian的算法相比，该算法可将内存消耗降低约95％。