Input File Generator

Fill in the parameters below to generate a ready-to-run Python input script for TENSO. The preview updates in real time, and warnings will appear if the chosen parameters may cause excessive memory usage or long runtimes.

New to TENSO? Read Getting Started first for a conceptual overview of each parameter. See Background for the theory behind HEOM and tensor networks.

Simulation Method

Propagation Method

Choose how the open-system dynamics are propagated. The form below adapts to the method you select.

Bath Parameters

Bath Correlation Function
Padé is generally more efficient at low temperatures.
Low-temperature correction terms. Each adds one bath mode to the hierarchy.
Drude–Lorentz Components

Each component contributes one bath mode. Spectral density: $J_\textrm{DL}(\omega) = \frac{2\lambda}{\pi}\frac{\omega_c\,\omega}{\omega^2+\omega_c^2}$

Brownian Oscillator Components

Each component contributes two bath modes. Spectral density: $J_\textrm{BO}(\omega) = \frac{4\lambda}{\pi}\frac{\eta\,\omega_0^2\,\omega}{(\omega^2-\omega_0^2)^2+4\eta^2\omega^2}$

System Parameters

Quantum System
Each ADM in the HEOM hierarchy is an $M\times M$ matrix.
System Hamiltonian sys_ham

$H_S$ — real or complex entries, units cm⁻¹

System–Bath Couplings sys_op(s)

Each coupling is an $M\times M$ operator $Q_n$ that couples the system to its own (independent) copy of the bath above. Add more than one for multi-bath / non-commuting fluctuations (e.g. one local projector per site).

Initial State
For HEOM the script builds ρ₀ = |ψ⟩⟨ψ| automatically; for MCTDH the wavefunction is used directly.

$|\psi\rangle$ — $M$ complex amplitudes (need not be normalized; conventions follow the basis order).

Time-Dependent Drive

Time-Dependent Drive optional

The drive adds a time-dependent term to the system Hamiltonian: $H(t) = H_S + f(t)\,V$, entering the dynamics as $-\tfrac{i}{\hbar}\,f(t)\,[V,\,\cdot\,]$. You define the scalar field $f(t)$ (returned in cm⁻¹, with $t$ in fs) and the operator $V$ it couples to.

$f(t) = A\,\exp\!\left[-\tfrac{1}{2}\left(\tfrac{t-t_0}{\sigma}\right)^2\right]\cos(\omega t + \varphi)$, with $\sigma = \mathrm{FWHM}/(2\sqrt{2\ln 2})$ and $\omega$ obtained from the carrier frequency.

Set equal to the system energy gap for a resonant pulse.
Drive operator td_op

$V$ — the $M\times M$ operator the field couples to (default $\sigma_x$ for a two-level system).

Drive preview f(t) over the propagation window
Live preview is unavailable for custom expressions — check the generated code below.

Solver & Tensor Network

Solver & Tensor Network
Truncation depth $N_k$. Increase until results converge (typical: 10–40).
Maximum rank of the auxiliary density matrices. Higher → more accuracy, more memory (typical: 32–128).
Absolute tolerance for the ODE integrator. Tighter → more accurate but slower.
Relative tolerance for the ODE integrator. Typical: 1e-5 to 1e-7.

Propagation Settings

Time Propagation
Outputs: {fname}.dat.log (density matrix) and {fname}.debug.log.

Advanced Settings

Advanced Settings optional

Expert controls. Each is emitted into the generated script only when you change it from its default, so the output stays clean. Options that don't apply to the chosen method are hidden.

Tensor-network shape. train is lean on memory; trees balance depth.
Starting bond rank of the tensor-network state.
Only used when DVR basis is enabled below.
Toggles

Estimated Resource Usage

Bath Modes (K)
Hierarchy Nodes
Est. RAM
Time Steps

Preview & Download


        
Getting Started API Reference