BIE::PFSimulation Class Reference

Sequential Monte Carlo Particle Filter sampler. More...

#include <PFSimulation.h>

Inheritance diagram for BIE::PFSimulation:
Collaboration diagram for BIE::PFSimulation:

List of all members.

Public Member Functions

 PFSimulation (int nsteps, int npart, Simulation *sim, Ensemble *factory)
 Constructor.
 ~PFSimulation ()
 Destructor.
void Run ()
 Run the actual simulation (without threads).
void Resume ()
 Resume restored simulation (without threads).
void ResumeLog ()
 Rollback state log file on resume.
void Nsample (int n)
 Change default number MCMC steps per resample (default=1).
void Restart ()
 For restarting simulation.
void SetTessTool (TessToolSender *tt)
 Set the TessTool communicator.
void SampleNext ()
 Request by TessTool to sample the Tile data.
void SwitchOnCLI ()
 Switch control to the CLI.
virtual void SuspendSimulation ()
 Request that the simulation suspend.
virtual void ResumeSimulation ()
 Request that the simulation resume.
virtual void SetAutoTessTool (int freq)
 Set the TessTool to update every.

Static Public Attributes

static double ESSprob
 ESS threshold for resampling (default: 0.1).
static int maxTries
 Maximum number of tries to find a good state to start (default: 400).

Friends

class boost::serialization::access


Detailed Description

Sequential Monte Carlo Particle Filter sampler.

Strategy

  1. Each of the N nodes maintains a fraction n=N_sample/N of the the data
  2. At the end of each step, the states of the n particles are swapped round robin for accumulating updates or likelihood computation

Currently, this will not work properly with TessTool.

The input data file should have randomized lines; e.g. a representative sample of the entire population for each subset is assumed.

The parameters nsteps may be used to constrain the simulation to a subset of the data. If nsteps=0, all data will be used.


Member Function Documentation

virtual void BIE::PFSimulation::SetAutoTessTool ( int  freq  )  [virtual]

Set the TessTool to update every.

Parameters:
freq iterations


The documentation for this class was generated from the following file:

Send suggestions, questions, and feedback to WEINBERG at ASTRO dot UMASS dot EDU.
Documentation generated at Fri Mar 26 00:35:12 2010 by doxygen