Template Class Hamiltonian_Dense

Inheritance Relationships

Base Type

Class Documentation

template<typename HilbertField>
class Hamiltonian_Dense : public Hamiltonian<HilbertField>

Public Functions

Hamiltonian_Dense(shared_ptr<model> the_model, const map<string, double> &value, sector _sec)

constructor

virtual void mult_add(vector<HilbertField> &x, vector<HilbertField> &y)

Applies the Hamiltonian: y = y +H.x.

Parameters:

  • y – vector to which H.x is added to

  • x – input vector

virtual void diag(vector<double> &d)

provides the diagonal d of H Used by the Davidson method

Parameters:

d – the diagonal of H (pre-allocated)

virtual double GS_energy()

Momber function for dense Hamiltionain that use LAPACK routine to find the lowest energy.

Parameters:

  • alpha – : first diagonal of the tridiagonal representation (returned by reference)

  • beta – : second diagonal of the tridiagonal representation (returned by reference) returns the GS energy.

virtual vector<shared_ptr<state<HilbertField>>> states(double &GS_energy)

Applies the Lanczos or Davidson-Liu algorithm for the lowest-energy states and energies.

returns a vector of pointers to states, to be treated by the parent model_instance.

Parameters:

GS_energy – : current ground-state energy of the model_instance, to be updated

virtual void print(ostream &fout)

Prints the dense form of the Hamiltonian Mostly for debugging purposes on very small systems.

matrix<HilbertField> dense_form()

Puts a dense form of the Hamiltonian matrix in the matrix H_dense Mostly for debugging purposes, but also for small dimensions.

virtual Q_matrix<HilbertField> build_Q_matrix(vector<vector<HilbertField>> &phi)

Constructs the Q_matrix (Lehmann representation) from full diagonalization.

Parameters:

phi – the initial vectors

Public Members

matrix<HilbertField> H_dense

dense form of the Hamiltonian (used when the dimension is small)