Reference.Files.Main

I/O files of the Matlab/Octave Version of WavePacket

Log files

After successfull termination of any of the WavePacket programs mentioned in this list there will be a logfile in the respective working directory. This file contains exactly the same information as displayed in the Matlab command window (Matlab console). This is achieved through the use of +prt/disp.m and/or +prt/warning.m and/or +prt/error.m. Examples:

• Quantum bound state calculations using qm_bound.m create a logfile named qm_bound.log

• Wavepacket or trajectory or etc propagations using qm_propa.m, create a logfile qm_propa.log

Movies and graphics files

Depending on the choice of different graphical representations, and depending on other user settings, some choice of the following graphics files will be produced upon running the programs qm_bound.m or qm_propa.m building on a position|momentum|phase space representation.

• abc_def.xyz with the following meaning

code	choices	explanation
abc	'wave' or 'traj' or ...	wave for wavefunctions represented on grids or traj for classical densities represented by trajectories or ..., see here for a complete list of main classes
def	'contour' or 'surface' or ...	Type of visualization chosen see here for a complete list of plot types
xyz	'mp4'	Animated graphics files (H.264, MPEG4) After having run qm_bound, the frames contain the sequence of eigenfunctions of the Hamiltonian (should be displayed step by step). After having run qm_propa, the frames contain snapshots of the time evolution (should be displayed continuously). Typically, the file size is of the order of 10 kB per frame.
xyz	'jpg'	Snapshot (JPEG graphics) of last time step
xyz	'fig'	Snapshot (Matlab figure file) of last time step

• wave_expect.jpg&fig or traj_expect.jpg&fig or etc: Curve plots showing populations and expectation values of energy (and other observables) vs. bound state index (qm_bound.m) or vs. simulated time (qm_propa.m). Produced only if plots.density is empty because otherwise expectation values are plotted alongside with (animated) densities

• wave_spectrum.jpg: Curve plots showing spectrum obtained by diagonalization of the Fourier grid Hamiltonian matrix (qm_bound.m) or obtained by Fourier transformation of the wavepacket autocorrelation function (qm_propa.m). Not available for trajectory-based simulations.

The additional (ODE based) program qm_optimal.m building on a state/energy representation (optionally) creates three more figures:

• uxy.fig: Evolution of input/control u(t), state vector x(t), output/observable y(t) in one figure with 2 or 3 subplots

• j12.fig: Evolution of target, cost, and total functionals in one figure with three subplots (qm_optimal.m only)

• psd.fig: Power spectral density of optimal field (qm_optimal.m only)

Many examples can be found in the demo section of the general WavePacket wiki.

Data files

See also the flow chart attached to our list of WavePacket programs

• Depending on the choice of the constituents of the underlying Hamiltonian, some operators, such as potential energy, dipole moment, or the initial wavefunctions may be read from (formatted!) data files. The respective tabulated values are then interpolated, see our detailed description.

• Stationary or time-dependent wavefunctions or trajectory bundles or etc (along with additional parameters) can (optionally!) be saved in unformatted Matlab data files wave_n.mat or traj_n.mat, see also our detailed description. This saving is done (optionally!) in qm_bound, qm_propa (as well as in qm_rerun and qm_regenerate). For the case of wavefunctions, these data can be loaded into qm_matrix, where they are used for calculation of matrix elements, or in qm_movie, where they are used to generate animated graphics (as well as in qm_rerun, qm_regenerate and qm_correlation).

• Matrix (eigen) representations of important operators (such as Hamiltonian, dipole, system-bath coupling) are calculated by qm_matrix.m and saved in files named wave_0.mat. They are loaded by qm_abncd where they are used to calculated matrices A,B,N,C,D and vectors x_i, x_e etc used for bilinear control problems.

• Matrices A,B,N,C,D and vectors x_i, x_e are generated by qm_abncd.m and saved in files named ket_0.mat or rho_0.mat for control problems in closed or open systems quantum dynamics, respectively.

• If desired, these matrices may be balanced by qm_balance.m after which they are saved in files named ket_bal_0.mat or rho_bal_0.mat.

• Subsequently, the balanced matrices may be truncated by qm_truncate.m after which they are saved in files named ket_xn_0.mat or rho_xn_0.mat where n is the (integer) dimensionality of the reduced system and where x can be either one of t (simple truncation) or s (singular perturbation theory).

• Alternatively, H2 optimal model reduction by qm_H2model.m can be used to generate reduced models saved in files named ket_hn_0.mat or rho_hn_0.mat where again n is the dimensionality of the reduced system.

• Any of the above matrices ket...mat or rho...mat can serve as main input to the program qm_optimal where the associated bilinear (optimal) control problem will be solved. The resulting solutions (time dependence of u,x,y-vectors and related quantities) are saved in files named ket..._control.mat or rho..._optimal.mat.

• After running an optimal control simulations using qm_optimal, the time-dependence of the optimal field can be found in (ASCII) formatted data file ket..._optimal_i.dat or rho..._optimal_i.dat where i stands for the number of backward-forward propagations after which the iteration has converged. These files can e.g. be used as input for propagations in the ODE setting (class ket or class rho) or also in the PDE setting (class wave) of qm_propa.