A JuliaQuantum package for solving dynamical equations in quantum mechanics.
The package is aimed at providing a unified framework for solvers for solving Dynamical Equations in Quantum Mechanics. Various solvers have been integrated from packages like ODE.jl, ExpmV.jl, Expokit.jl, using base features from QuBase.jl. QuBase.jl is the base package which includes most of the basic type constructs and methods used in Quantum Mechanics.
The following Dynamical Equations can be solved using this package :
- Schrodinger Equation
- Liouville von Neumann Equation
- Lindblad Master Equation
Using the following solvers :
-
Euler Method, Crank-Nicholson Method, Krylov subspace Method
-
ode45, ode78, ode23s which have been integrated from ODE.jl
-
Two versions of
expmv
implementations from :a. ExpmV.jl b. Expokit.jl
-
Quantum Monte-Carlo Wave Function Method has been integrated with reference from the article : Monte Carlo Simulation of the Atomic Master Equation for Spontaneous Emission, Phys. Rev. A 45, 4879 (1992) by R. Dum, P. Zoller, and H. Ritsch.
As the solvers have been integrated from various packages and also that QuDynamics is based on QuBase.jl, we need to install the following for using QuDynamics.
Dependencies
- QuBase.jl
This has to be added using
Pkg.clone("https://github.com/JuliaQuantum/QuBase.jl.git")
- Expokit.jl
This has to be added using
Pkg.clone("https://github.com/acroy/Expokit.jl.git")
- ExpmV.jl
This has to be added using
Pkg.clone("https://github.com/marcusps/ExpmV.jl.git")
The current package can be used by
Pkg.clone("https://github.com/JuliaQuantum/QuDynamics.jl.git")
We could start using the package by :
using QuBase
using QuDynamics
QuPropagator
is the starting point in order to compute or plot, related entities like states at different time steps, expectation values of an observable and so on. QuProagator
is the one which takes
in the configuration of the system along with the solver which is to be used for solving.
For example, consider the following configuration :
hamiltonian = sigmax # the Hamiltonian of the system
initial_state = statevec(1, FiniteBasis(2)) # the initial state of the system
tlist = 0.:0.1:2*pi # the time steps at each of which we have the evolved state
Now we have the basic ingredients to construct the QuPropagator
, assuming we have a solver in mind.
We construct the following :
qprop_ode45 = QuPropagator(hamiltonian, initial_state, tlist, QuODE45())
The general format of QuPropagator
construct is as follows :
QuPropagator(Hamiltonian_of_the_system, Collapse_operators, Initial_state, Time_step_array, Solver)
Alternatively we can also pass :
QuPropagator(Type_of_Equation, Initial_state, Time_step_array, Solver)
Implying, we could do the following which is equivalent to the above construction of qprop_ode45
schrodinger_eq = QuSchrodingerEq(hamiltonian)
qprop_ode45 = QuPropagator(schrodinger_eq, initial_state, tlist, QuODE45())
A common convention for the equations
and solvers
types and methods is they start with a prefix Qu
.
For example :
`Schrodinger Equation` => `QuSchrodingerEq`
`ode45` => `QuODE45`
`expmv` => `QuExpmV` or `QuExpokit` (based on the packages)
Now we are all set to calculate the states at various time steps which can be achieved in the following ways :
Method - 1: To get the evolved state at every time instant, we can do the following :
for (t, psi) in qprop_ode45
println("t : $t, psi : $psi")
end
Method - 2: To get the evolved state at the next time instant (here we get the state after the fist time-step), we can do the following :
next_state = propagate(QuODE45(), QuSchrodingerEq(h), t[2], t[1], initial_state)
println(next_state)
For more examples and work cases, please refer to the examples folder of the repo.
Note : Instances of QuPropagator
can also be called by QuStateEvolution
. Also the propagation of evolution operator can be
evaluated using QuEvolutionOp
.
The following discussions tend to provide an insight on the design of the internals :
- The GSoC blog updates
- Design and implementation related discussion a, b
- Issues and Pull Requests are also good references which shed light on the development of the package.
- Wiki page has a mention of the design and implementation related discussion.
We would love to have contributors. The process is simple :
- Fork and clone the repo.
- Create a local branch and implement the changes.
- Test the changes and send in a pull request.