This was the first version put onto the anonymous ftp server
medphys.ucl.ac.uk
Version 1.01 30th Sept 1994.
This corrected a number of problems with the original code, so that it now runs
properly on the acc (Sun ANSI C) and gcc (GNU C/C++) compilers. The original
version only compiled properly only on g++ (GNU C++) and CC (Sun C++)
compilers. Some of my code was obviosly C++, when I though it was ANSI C!
Version 1.02 14 Oct 1994
This release corrected a number of memory allocation errors in 'yagi', 'output'
and 'optimise'. Memory was not always freed in earlier releases. Also, a for
loop in all 3 programs had a calls to routines which allocated memory many
times, when it was only needed once. The output file now has # to mark comments,
which the gnuplot (GNU plotting program) recognises.
Version 1.03
'optimise' was compleatly re-written, so that it no longer used lots of
tempory files on disk. This makes it much faster (6-7x on a 6ele beam ).
Version 1.04 First DOS based version.
Version 1.05 26 Oct 1994.
Optimise now takes lots of options. You dont need such a complex command to get it to work. There are sensible defaults. Some options dont work on optimise - '-g' , '-T' , -t' and '-p'.
Version 1.06 28/10/94
Released on to one user. Options 'T' and 't' now work. Genetic algorithm still does, although man page mention them.
Version 1.07 31/10/94
Added option '-w' that averages data of 3 frequencies to build wide-band antenna. A lot of integers for best/latest design were removed, puting them instead in a structure.
Version 1.08 5/11/94.
Genetic algorithm works now. Parameters used internally (probability of crossover etc) may not be anywhere near optimum. GA does not seem to offer any real advantage, but early days yet.
Added a new program 'rantest' to check the randome number generators are working
okay.
Version 1.09
Added a facility in 'output' to print 3dB bandwidths. Removed printing of parameter 'F', as this is of little use and space is needed.
Changed default elements altering form only parasitic to all elements.
Corrected an error in which the .gai file prints the E-plane data and H-plane data refersed.
Version 1.10 Thursday 3 October 1995.
Corrected a bug where 'output' would fail, due to a problem with assuming
where the 3dB points were in the H plane. Added options '-i' so you can
completely ignore 3dB beamwidths, also options '-E' and '-H' so you
can specify the maximum value of theta and phi over which the beamwidths
are calculated. A better fix, wherby the program always gets it right
will possibly be added, but this will slow the program considerably. Needs
some though of the best way to do this.
Program now computes the maximum level of any sidelobe. Obviously
there no point worrying about getting a good FB, if one of the other
lobes is more significant.
Version 1.11 Never released.
Changed most of the things done in 1.10! The '-i' option on optimise has been droped, but insread a '-e' option was added to avoid calculation of the E-palne
BW and a -h to avoid calculation of the H-plane BW.
Version 1.12
Changed GA length. eg min length can now go below 0.4 lambda. Changed
program 'first; to calculate the according to the well known DL6WU antennnas.
fitness computed in a different was. weight.swr=1-log10(swr), rather than the
old 1-((swr-1)/(swr+1))^2.*weight.gain.
'yagi' now prints the element currents, as a row of *'s, if the -d option is