Utilities

jVMC.util.util.ground_state_search(psi, ham, tdvpEquation, sampler, numSteps=200, varianceTol=1e-10, stepSize=0.01, observables=None, outp=None)

This function performs a ground state search by Stochastic Reconfiguration.

Arguments:

• psi: Variational wave function (jVMC.vqs.NQS)

• ham: Hamiltonian operator

• tdvpEquation: An instance of jVMC.util.TDVP or jVMC.util.MinSR

• numSteps: Maximal number of steps

• varianceTol: Stopping criterion

• stepSize: Update step size (learning rate)

• observables: Observables to be measured during ground state search

• outp: None or instance of jVMC.util.OutputManager.

jVMC.util.util.measure(observables, psi, sampler, numSamples=None)

This function measures expectation values of a given set of operators given a pure state.

Arguments:

• observables: Dictionary of the form with operator names as keys and (lists of) operators as values, e.g.:

  { "op_name_1": [operator1, operator2], "op_name_2": operator3 }
  

  If an operator takes additional positional arguments, a tuple of the operator together with the arguments has to be passed instead. For example, assuming that operator3 from above requires additional arguments *args:

  { "op_name_1": [operator1, operator2], "op_name_2": [(operator3, *args)] }
  

• psi: Variational wave function (instance of jVMC.vqs.NQS)

• sampler: Instance of jVMC.sampler used for sampling.

• numSamples: Number of samples (optional)

Returns:

A dictionary holding expectation values, variances, and MC error estimates for each operator. E.g. for the exemplary operator input given in Arguments:

{ 
    "op_name_1": { "mean": [mean1, mean2],
                   "variance": [var1, var2],
                   "MC_error": [err1, err2] },
    "op_name_2": { "mean": [mean3],
                   "variance": [var3],
                   "MC_error": [err3] }
}