4. METHOD OF SOLUTION
The error function and complementary error
function are approximated by rational functions. Three such
rational approximations are used depending on whether -
^x^.LT.0.46875,
0.46875.LE.^x^.LT.4.0, or
^x^.GE.4.0.
In the first region the error function is computed directly and
the complementary error function is computed via the identity
erfc(x)=1.0-erf(x). In the other two regions the complementary
error function is computed directly and the error function is
computed from the identity erf(x)=1.0-erfc(x). The error function
and complementary error function are real-valued functions of any
real argument. The range of the error function is (-1,1). The
range of the complementary error functon is (0,2).