The QDesignOptimizer

Tuning up a superconducting quantum chip often involves lenghty simulations where the user manually updates the design variables, such as capacitance lengths and Josephson junction sizes, in order to reach the target quantities of the circuit, such as frequencies, decay rates and coupling strengths. The QDesignOptimizer package strongly reduces the need for manual intervention by automating the simulation and optimization of the design variables. The optimization cycle first runs a detailed electro-magnetic HFSS simulation combined with EPR analysis to estimate the current values of the circuit quantities integrated together with design framework qiskit-metal. The second step solves a nonlinear approximate model based on the user’s physical knowledge about the circuit, which estimates how the design variables should be changed to reach the target. The flexibility in the QDesignOptimizer setup allows for efficient investigation of chip subsystems, since the simulation and approximate physical model are dynamically compiled from the user settings.

Reference

A technical paper describing the functionality and giving examples is in preparation.
Eriksson, et al., “Automatic multi-parameter design optimization for superconducting devices”, in preparation.
Please refer to this publication when citing the QDesignOptimizer.

Code & documentation

The code and documentation are available from github and also pip-installable via qdesignoptimizer.

Events

Kick-off webinar

We are organizing a webinar to help researcher getting started with their own quantum design optimization. The webinar will take place on zoom on April 25, 2025, 3-6pm CEST (Amsterdam time). Please sign up via this link to receive further program information.