.. _gallery_optimization:

Optimization
============

Optiland supports optimization through different backends.  
The **NumPy backend** relies on SciPy optimizers, while the **Torch backend**  
uses PyTorch's native optimization tools.  

NumPy (SciPy) Optimization
--------------------------

These examples show how Optiland integrates with SciPy's optimizers  
to improve optical systems using classical numerical methods.  

.. nbgallery::
    optimization/rms_spot_size
    optimization/wavefront_error
    optimization/undo
    optimization/pickups
    optimization/constrained
    optimization/bounded_operands
    optimization/orthogonal_descent
    optimization/global
    optimization/basin_hopping
    optimization/shgo
    optimization/custom_scaler
    optimization/nurbs_freeform_telescope

Torch Optimization
------------------

These examples demonstrate optimization using the Torch backend,  
taking advantage of PyTorch's autograd and optimizers for differentiable design.  

.. nbgallery::
    optimization/torch_rms_spot_size
    optimization/torch_constrained
    optimization/torch_module_rms_spot
    optimization/torch_module_custom_objective