| Author |
Comment/Response |
Dave
|
 10/29/08 1:52pm
Hello,
I've created the following auxiliary function which outputs some data to a file, calls an external FORTRAN program which reads the data. Mathematica then reads the output from the FORTRAN code, and then performs a calculation, a least squares sum.
Lse[x_,y_]:=(
PSiH[[2,1]]=x;
PSiSi[[1,2]]=y;
Export["PSiH.dat",PSiH];
Export["PSiSi.dat",PSiSi];
in=ReadList["!sirebo_cygwin.exe<mat.in",Number, RecordLists->True];
ESiH=Extract[in[[2]],3];
ESiSi=Extract[in[[3]],3];
((ESiH-(-3.124))^2+(ESiSi-(-1.731))^2)^(1/2)
)
By itself the function works fine, if I give it a value of say Lse[1,1], I get a reasonable output. So far so good. However, I would like to use a standard Mathematica routine like "Minimize" to find optimal values for the input parameters x and y to minimize the least squares sum in the last line.
I tried:
Minimize[Lse[x,y],{x,y}]
that gives an error. It actually writes "x" and "y" to the data files which crashes the FORTRAN code.
Any suggestions how to get around this?
Thanks!
URL: , |
|
