dolphindes 0.2.1

A Python package for calculating limits on the performance of photonic devices using dual optimization methods.

Dolphindes 🐬

Dolphindes (very loosely, Dual Optimization Limits for PHotonic/PHysical INverse DESign) is a Python package for calculating limits on the performance of photonic devices using dual optimization methods. It can calculate structure-agnostic performance bounds for a wide range of photonic problems. The package works by relaxing the photonic inverse design problem into a field optimization problem, which can then be further relaxed into a convex problem using Lagrange duality.

📦 Installation

1. Install System Dependencies

Make sure you have libsuitesparse-dev installed. This is required by scikit-sparse

For Debian/Ubuntu systems:

sudo apt-get update
sudo apt-get install libsuitesparse-dev

2. Clone this repo and activate the provided conda environment dolphindes.yml

3. If using your own environment, instead run

pip install .

🔧 Running Tests

To run the dolphindes tests, simply run

pytest

Optionally, provide the -s flag to print the output of the tests. You will need to have pytest and pytest-dependency installed in your environment.

📚 Documentation and Tutorials

Documentation may be found at dolphindes.readthedocs.io

Citations

If you use dolphindes in your work, please cite the following paper:

[Review article coming soon]

Additionally, if you use dolphindes to do Verlan design, you should cite the initial Verlan papers:

@article{chao_amaolo_blueprints_2025,
      title = {Bounds as blueprints: towards optimal and accelerated photonic inverse design},
      author = {Pengning Chao and Alessio Amaolo and Sean Molesky and Alejandro W. Rodriguez},
      journal = {Opt. Express},
      keywords = {Fourier transforms; Inverse design; Raman scattering; Ring resonators; Stochastic processes; Whispering gallery modes},
      number = {5},
      pages = {7337--7350},
      publisher = {Optica Publishing Group},
      volume = {34},
      month = {Mar},
      year = {2026},
      url = {https://opg.optica.org/oe/abstract.cfm?URI=oe-34-5-7337},
      doi = {10.1364/OE.585505},
}

@article{molesky_verlan_2025,
    title = {Inferring {{Structure}} via {{Duality}} for {{Photonic Inverse Design}}},
    author = {Molesky, Sean and Chao, Pengning and Amaolo, Alessio and Rodriguez, Alejandro W.},
    year = {2025},
    month = apr,
    number = {arXiv:2504.14083},
    eprint = {2504.14083},
    primaryclass = {math},
    publisher = {arXiv},
    doi = {10.48550/arXiv.2504.14083},
    archiveprefix = {arXiv}
}