Propagator computation for dynamics #3437
Description
Required prerequisites
- Search the issue tracker to check if your feature has already been mentioned or rejected in other issues.
Describe the feature
As part of my workflow, I need to convert a time independent operator into the interaction frame with respect to a general time dependent Hamiltonian (which doesn't necessarily commute with itself at different times). It would be great if I could do this on the dynamics backend as efficiently as possible, using GPUs.
One way to do this that is more generally applicable would be to introduce a function which calculates the propagator of a time dependent Hamiltonian via a time-ordered exponential. This would be equivalent to QuTiP's propagator method. This is a general function which would be useful in a wide range of applications, and would be great to have it built efficiently with CUDA.
Are there any plans to implement something like this soon?