openmmslicer.sampling_metrics module¶

This module contains the EnergyCorrelation metric, which can be used for adaptively determining the number of decorrelation steps at each SMC iteration.

class openmmslicer.sampling_metrics.EnergyCorrelation(ensemble)¶

Bases: object

A sampling metric which determines the energy autocorrelation time and stops sampling after enough decorrelation.

Parameters: ensemble (slicer.smc.SequentialSampler) – A sampler, containing all of the information of the system.

ensemble¶

The associated SMC sampler.

Type: slicer.smc.SequentialSampler

E_before¶

A list of all reduced initial potential energies for each walker.

Type: [float]

E_after¶

A list of all reduced current potential energies for each walker.

Type: [float]

evaluateAfter()¶: Evaluates the energies at the current timestep.

evaluateBefore()¶: Evaluates the energies immediately after resampling.

property max¶

The maximum value of the metric. Default is 0.1.

Type: float

property metric¶

The energy autocorrelation coefficient.

Type: float

property min¶

The minimum value of the metric. Default is -0.1.

Type: float

property requireNextLambda¶

Whether a potentially expensive calculation of the next lambda value is needed.

Type: bool

reset()¶: Resets the sampling metric.

serialise()¶

property terminateSampling¶

Whether the metric is within the desired bounds.

Type: bool