Lindblad steady state
Lindblad steady state
I am quite new to MPS, but I was wondering if your python library could be used to calculate the steady state of a Lindbladian by e.g. time evolution, since in the documentation, if I didn't miss anything, only hermitian cases are discussed.
Re: Lindblad steady state
Unfortunately, we don't have code for doing this in TeNPy as of now.
There had been some work in that direction, see this topic, but I didn't hear back since that initial zoom meeting.
If you want to implement something in that direction, that would be very welcome
There had been some work in that direction, see this topic, but I didn't hear back since that initial zoom meeting.
If you want to implement something in that direction, that would be very welcome
Re: Lindblad steady state
Hey sorry somehow i didnt get any mail notification about your answer so i assumed there was none!
Thanks so much for the reply!
So there has been an implementation using MPS by someone in my group some time ago also using ITensor.
https://journals.aps.org/prb/pdf/10.110 ... 101.165132
Yeah so since this library you offer seems very convenient to use and extend i would indeed like to work on that.
So I am curious about something -> To my understanding, and correct me if i am wrong, there should be two options -> Either modify the lanczos part of the TDVP solver to bilanczos, this i guess should then suffice?
For TEBD one has to use the nearest neighbour model, and it emphasizes that it needs nearest neighbour in MPS sense, however using SpinHalfFermionSite still works even though two equal spin sites are not nearest neighbour in MPS sense right? So since I use the so called super fermion representation what would suffice for me is this:
Creating a "Site" which allows for two spin up and two spin down. Like for spin up and down each of the four (lets distinguish them like spin up/down normal/tilde) can jump to a neighbouring site of its kind, however, within the site a jump between "normal" and "tilde" are also allowed.
Would this work?
Looking forward to an answer and would also be happy to discuss this on videocall or smth if you think its interesting!
Thanks so much for the reply!
So there has been an implementation using MPS by someone in my group some time ago also using ITensor.
https://journals.aps.org/prb/pdf/10.110 ... 101.165132
Yeah so since this library you offer seems very convenient to use and extend i would indeed like to work on that.
So I am curious about something -> To my understanding, and correct me if i am wrong, there should be two options -> Either modify the lanczos part of the TDVP solver to bilanczos, this i guess should then suffice?
For TEBD one has to use the nearest neighbour model, and it emphasizes that it needs nearest neighbour in MPS sense, however using SpinHalfFermionSite still works even though two equal spin sites are not nearest neighbour in MPS sense right? So since I use the so called super fermion representation what would suffice for me is this:
Creating a "Site" which allows for two spin up and two spin down. Like for spin up and down each of the four (lets distinguish them like spin up/down normal/tilde) can jump to a neighbouring site of its kind, however, within the site a jump between "normal" and "tilde" are also allowed.
Would this work?
Looking forward to an answer and would also be happy to discuss this on videocall or smth if you think its interesting!