Bayesian Analysis of Neutron Star Mass and Radius Observations
Public Member Functions | Protected Attributes
bamr::bamr_class Class Reference

Statistical analysis of EOS from M and R constraints. More...

Detailed Description

Todo:
Right now the EOS is rejected if the pressure decreases at any density, when in reality, it should only check if the pressure decreases at a density below that of the central density of the maximum mass star. This isn't a problem for current EOSs, but could be a problem in the future.
Todo:
It's not clear if successive calls of the mcmc command really work. For now, one may have ensure the program exits after each mcmc() run.
Todo:
Fix issue of block_counter giving confusing output if there are too few MCMC points between each block.
Todo:
More testing
Todo:
Better documentation
Todo:
Help with plots
Idea for Future:
Allow non-tabulated data specified as a function?

Definition at line 79 of file bamr.h.

#include <bamr.h>

Public Member Functions

virtual void run (int argc, char *argv[])
 Main wrapper for parsing command-line arguments.
 

Protected Member Functions

Main functions called from the command-line interface
virtual int set_model (std::vector< std::string > &sv, bool itive_com)
 Set the model for the EOS to use.
 
virtual int set_first_point (std::vector< std::string > &sv, bool itive_com)
 Set the first point in the parameter space.
 
virtual int add_data (std::vector< std::string > &sv, bool itive_com)
 Add a data distribution to the list.
 
virtual int mcmc (std::vector< std::string > &sv, bool itive_com)
 Run a MCMC simulation. More...
 

Protected Attributes

o2scl::err_hnd_cpp ee
 Error handler for each thread.
 
o2scl::table_units tc
 Main data table for Markov chain.
 
size_t mh_success
 The number of Metropolis steps which succeeded.
 
size_t mh_failure
 The number of Metropolis steps which failed.
 
size_t mcmc_iterations
 Total number of mcmc iterations.
 
std::vector< double > markov_chain
 Store the full Markov chain.
 
o2scl::cli cl
 Command-line interface.
 
bool first_file_update
 If true, then first_update() has been called.
 
int grid_size
 Number of bins for all histograms (default 100)
 
size_t n_chains
 Number of Markov chain segments.
 
size_t chain_size
 Number of chains.
 
std::string model_type
 A string indicating which model is used, set in set_model().
 
bool has_esym
 True if the model provides S and L.
 
bool has_eos
 True if the model has an EOS.
 
size_t nparams
 Number of parameters, set in set_model()
 
double schwarz_km
 Schwarzchild radius (set in constructor)
 
model * modp
 The model for the EOS.
 
model * modp2
 The second copy of the model for the EOS.
 
o2scl::rng_gsl gr
 Random number generator.
 
std::ofstream scr_out
 The screen output file.
 
std::vector< std::string > run_args
 The arguments sent to the command-line.
 
Parameter objects for the 'set' command
o2scl::cli::parameter_double p_max_time
 
o2scl::cli::parameter_double p_min_max_mass
 
o2scl::cli::parameter_double p_step_fac
 
o2scl::cli::parameter_double p_exit_mass
 
o2scl::cli::parameter_double p_input_dist_thresh
 
o2scl::cli::parameter_double p_min_mass
 
o2scl::cli::parameter_int p_n_warm_up
 
o2scl::cli::parameter_int p_grid_size
 
o2scl::cli::parameter_int p_user_seed
 
o2scl::cli::parameter_int p_max_iters
 
o2scl::cli::parameter_int p_file_update_iters
 
o2scl::cli::parameter_bool p_norm_max
 
o2scl::cli::parameter_bool p_debug_star
 
o2scl::cli::parameter_bool p_debug_load
 
o2scl::cli::parameter_bool p_debug_eos
 
o2scl::cli::parameter_bool p_output_next
 
o2scl::cli::parameter_bool p_baryon_density
 
o2scl::cli::parameter_bool p_use_crust
 
o2scl::cli::parameter_bool p_inc_baryon_mass
 
o2scl::cli::parameter_double p_nb_low
 
o2scl::cli::parameter_double p_nb_high
 
o2scl::cli::parameter_double p_e_low
 
o2scl::cli::parameter_double p_e_high
 
o2scl::cli::parameter_double p_m_low
 
o2scl::cli::parameter_double p_m_high
 
Histogram limits
double nb_low
 
double nb_high
 
double e_low
 
double e_high
 
double m_low
 
double m_high
 
Grids
o2scl::uniform_grid< double > nb_grid
 
o2scl::uniform_grid< double > e_grid
 
o2scl::uniform_grid< double > m_grid
 
Other parameters accessed by 'set' and 'get'
int file_update_iters
 The number of MCMC successes between file updates.
 
bool debug_load
 If true, output debug information about the input data files (default false)
 
bool norm_max
 If true, normalize the data distributions so that the max is one, otherwise, normalize so that the integral is one.
 
int max_iters
 Maximum number of iterations (default 0)
 
bool use_crust
 If true, use the default crust (default true)
 
bool debug_star
 If true, output stellar properties for debugging.
 
bool debug_eos
 If true, output equation of state for debugging.
 
bool output_next
 If true, output next point.
 
bool baryon_density
 If true, compute the baryon density.
 
double step_fac
 MCMC stepsize factor (default 15)
 
double input_dist_thresh
 The lower threshold for the input distributions.
 
double exit_mass
 An upper mass threshold.
 
double min_mass
 Minimum mass allowed for any of the individual neutron stars.
 
double min_max_mass
 Minimum allowed maximum mass.
 
int n_warm_up
 Number of warm up steps (successful steps not iterations) More...
 
double max_time
 Time in seconds (3600 seconds is one hour, default is 86400 seconds or 1 day)
 
int user_seed
 If non-zero, use as the seed for the random number generator.
 
bool best_detail
 If true, output more detailed information about the best point.
 
bool inc_baryon_mass
 If true, output information about the baryon mass as well as the gravitational mass.
 
Limits on mass and radius from source data files

These are automatically computed in load_mc() as the smallest rectangle in the $ (M,R) $ plane which encloses all of the user-specified source data

double in_m_min
 
double in_m_max
 
double in_r_min
 
double in_r_max
 
MPI properties
int mpi_rank
 The MPI processor rank.
 
int mpi_nprocs
 The MPI number of processors.
 
double mpi_start_time
 The MPI starting time.
 
Input neutron star data
std::vector< o2scl::table3d > source_tables
 Input probability distributions.
 
std::vector< std::string > source_names
 The names for each source.
 
std::vector< std::string > table_names
 The names of the table in the data file.
 
std::vector< std::string > source_fnames
 File names for each source.
 
std::vector< std::string > slice_names
 Slice names for each source.
 
size_t nsources
 The number of sources.
 
std::vector< double > first_mass
 The initial set of neutron star masses.
 
Parameter limits
entry low
 
entry high
 
Other variables
ubvector first_point
 The first point in the parameter space.
 
std::string first_point_file
 The file containing the initial point.
 
TOV solver objects
o2scl::eos_tov_interp teos
 EOS interpolation object for TOV solver.
 
o2scl::tov_solve * ts
 Pointer to a TOV solver.
 
o2scl::tov_solve * ts2
 Second pointer to a TOV solver.
 
o2scl::tov_solve def_ts
 Default TOV solver.
 
o2scl::tov_solve def_ts2
 Second default TOV solver.
 

Integer designating how to set the initial point

int first_point_type
 
static const int fp_unspecified =-1
 
static const int fp_last =-2
 
static const int fp_best =-3
 

Internal functions

virtual void select_mass (entry &e_current, entry &e_next, double mmax)
 The function which selects the next neutron star masses.
 
virtual void table_names_units (std::string &s, std::string &u)
 Setup column names and units for data table.
 
virtual int mcmc_init ()
 Make any necessary preparations for the mcmc() function. More...
 
virtual bool make_step (double w_current, double w_next, bool debug, bool warm_up, int iteration)
 Decide to accept or reject the step.
 
virtual void init_grids_table (entry &low, entry &high)
 Initialize the expectation value objects. More...
 
virtual void prepare_eos (entry &e, model &modref, o2scl::tov_solve *tsr, int &success)
 Further preparations of the EOS before calling the TOV solver. More...
 
virtual void add_measurement (entry &e, o2scl::o2_shared_ptr< o2scl::table_units<> >::type tab_eos, o2scl::o2_shared_ptr< o2scl::table_units<> >::type tab_mvsr, double weight, bool new_meas, size_t n_meas, ubvector &weights)
 Add a measurement.
 
virtual void fill_line (entry &e, o2scl::o2_shared_ptr< o2scl::table_units<> >::type tab_eos, o2scl::o2_shared_ptr< o2scl::table_units<> >::type tab_mvsr, double weight, bool new_meas, size_t n_meas, ubvector &weights, std::vector< double > &line)
 Fill vector in line with data from the current Monte Carlo point.
 
virtual void first_update (o2scl_hdf::hdf_file &hf, model &modp)
 Write initial data to HDF file.
 
virtual void update_files (std::string fname_prefix, model &modp, entry &e_current)
 Write histogram data to files with prefix fname.
 
virtual void setup_cli ()
 Set up the 'cli' object. More...
 
virtual void load_mc ()
 Load input probability distributions (called by mcmc())
 
virtual double compute_weight (entry &e, model &modref, o2scl::tov_solve *ts, int &success, ubvector &wgts, bool warm_up)
 Compute $ P(M|D) $ from parameters in entry e. More...
 
virtual void output_best (std::string fname_prefix, entry &e_best, double w_best, o2scl::o2_shared_ptr< o2scl::table_units<> >::type tab_eos, o2scl::o2_shared_ptr< o2scl::table_units<> >::type tab_mvsr, ubvector &wgts)
 Output the "best" EOS obtained so far (called by mcmc())
 
virtual void compute_star (entry &e, model &modref, o2scl::tov_solve *ts, int &success)
 Tabulate EOS and then use in cold_nstar. More...
 

Return codes for each point

std::vector< int > ret_codes
 
static const int ix_success =0
 
static const int ix_mr_outside =1
 
static const int ix_r_outside =2
 
static const int ix_zero_wgt =3
 
static const int ix_press_dec =4
 
static const int ix_nb_problem =5
 
static const int ix_nb_problem2 =6
 
static const int ix_crust_unstable =7
 
static const int ix_mvsr_failed =8
 
static const int ix_tov_failure =9
 
static const int ix_small_max =10
 
static const int ix_tov_conv =11
 
static const int ix_mvsr_table =12
 
static const int ix_acausal =13
 
static const int ix_acausal_mr =14
 
static const int ix_param_mismatch =15
 
static const int ix_neg_pressure =16
 
static const int ix_no_eos_table =17
 
static const int ix_eos_solve_failed =18
 

Member Function Documentation

virtual int bamr::bamr_class::mcmc ( std::vector< std::string > &  sv,
bool  itive_com 
)
protectedvirtual

In order to save the code from copying results over if a step is rejected, the Metropolis steps alternate between two kinds. In the first kind of step, modp is the previous point and modp2 is the new point to be considered, and in the second kind of step, modp2 is the old point and modp is the new point to be considered. This two-step procedure is also why there are two TOV solvers, i.e. ts and ts2.

virtual int bamr::bamr_class::mcmc_init ( )
protectedvirtual

This is called by mcmc(). If the return value is non-zero then it is assumed that the calculation fails and mcmc() returns.

This function does nothing and returns zero by default.

virtual void bamr::bamr_class::init_grids_table ( entry &  low,
entry &  high 
)
protectedvirtual

This function is called by mcmc().

virtual void bamr::bamr_class::prepare_eos ( entry &  e,
model &  modref,
o2scl::tov_solve *  tsr,
int &  success 
)
protectedvirtual

This function is empty by default.

virtual void bamr::bamr_class::setup_cli ( )
protectedvirtual

This function just adds the four commands and the 'set' parameters

virtual double bamr::bamr_class::compute_weight ( entry &  e,
model &  modref,
o2scl::tov_solve *  ts,
int &  success,
ubvector &  wgts,
bool  warm_up 
)
protectedvirtual

Called by mcmc().

virtual void bamr::bamr_class::compute_star ( entry &  e,
model &  modref,
o2scl::tov_solve *  ts,
int &  success 
)
virtual

Called by compute_weight().

Todo:
Temporarily made public for drdp project hack

Field Documentation

int bamr::bamr_class::n_warm_up
protected
Note
Not to be confused with warm_up, which is a boolean local variable in some functions not an int.

Definition at line 188 of file bamr.h.

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

Documentation generated with Doxygen. Bamr documentation is under the GNU Free Documentation License.