lundsga@stolaf.UUCP (Soren K. Lundsgaard) (04/06/84)
Does anyone have a fitting algorithm for the neutron experiment that *everyone* has done in college. This is the experiment where a piece of silver is exposed to a neutron source, then is taken out of the source, and counts of its activity are taken over intervals of time. The equation that this should fit to is a1 * exp( a2 * t ) + a3 * exp( a4 * t ) + a5 where a1 is the coefficient for one of the silver isotopes, a3 is the coefficient for the other isotope, a2 and a4 are the rate of decay corresponding to the half lives of the isotopes, and a5 is the background radiation. Please respond by mail. Soren K. Lundsgaard, St. Olaf College, Northfield MN. ihnp4!stolaf!lundsga
gwyn@brl-vgr.ARPA (Doug Gwyn ) (04/08/84)
As a grad student in the Physics Dept. at Clark Univ. in Worcester, MA, I programmed (in portable ANS 66 Fortran) a general non-linear least- squares curve fitting algorithm, using basically a Marquardt gradient- expansion algorithm. This program was in production use to fit noisy PAC & Mossbauer data (errors were used for weighting). This would certainly fit the requirements for fitting your double-exponential data from the silver experiment. You should contact the Clark Phys. Dept. to see about getting this code. (Bob Suter is a good person to talk to.)