devries / analytic_infall Goto Github PK

Thank you for downloading the analytic infall models. These models are
discussed in detail in De Vries & Myers (2005). More documentation and updates
are available from http://cfa-www.harvard.edu/~cdevries/analytic_infall.html
. Although this paper includes all the analytic models discussed in the paper,
I would restrict myself to the HILL5 and TWOLAYER6 models, which were found to
be the best models. 

In order to fit a spectrum you need to place it in a two column text format
(the first column should be velocity in kilmeters per second, and the second
should be brightness temperature in Kelvins). Lines that begin with a hash mark
'#' will be ignored. Use the hybrid version of the model you want to run, this
will allow for some slow differential evolution minimization which will
(hopefully) find the global minimum well followed by Nelder-Mead simplex
minimization which will quickly find the bottom of that well. The hybrid
programs of interest are "hill5_hybrid" and "twolayer6_hybrid". The arguments
to both programs are:

1. input filename  --- the file with your spectrum
2. frequency --- to perform the J(T) conversion. A frequency of 0 will work in
		 units of J(T) instead of T.
3. vmin --- minimum velocity of the line profile.
4. vmax --- maximum velocity of the line profile. Be sure to get the entire
	    line full width as well as a little baseline. This is the region
	    over which chi-squared is calculated, so artifacts of the fit can
	    sometimes appear outside this region.
5. population in generation --- The number of solutions to calculate each
				generation of the differential evolution. I
				usually use about 200-300 here.
6. generations per check --- The number of generations to run before checking
			     for convergence in the differential evolution. I
			     usually pick 300 here.
7. checks to convergence --- The number of checks to make before deciding the
			     differential evolution algorithm has converged. I
			     usually pick 3 to 5 here. 
8. output file --- Where to place the fit and parameters.

The output file will have a header line the following (a HILL 5 example is
shown):

# Tau:   4.46394
# Vlsr:  0.000296732
# Vin:   0.0993918
# sigma: 0.0937655
# Tpeak: 10.5371
# Attained Chisq: 0.539567

giving the parameters of the fit and Chi Squared (not a reduced
chi-squared). Followed by the fit in a three column data file. The first two
columns are the two columns of the input file, while the third column is the
fit of the analytic model. Always check that the fit is good.

Please email me at [email protected] if you have any questions or
comments.