I’ve started a new project on Github for a new modern object-oriented version of the Pikaia genetic algorithm code. This is a refactoring and upgrade of the Pikaia code from the High Altitude Observatory. The original code is public domain and was written by Paul Charbonneau & Barry Knapp (version 1.2 was released in 2003). The new code differs from the old code in the following respects:
- The original fixed-form source (FORTRAN 77) was converted to free-form source.
- The code is now object-oriented Fortran 2003/2008. All user interaction is now through a class, which is the only public entity in the module.
- All real variables are now 8 bytes (a.k.a. “double precision”). The precision can also be changed by the user by modifying the
- The random number generator was replaced with calls to the intrinsic Fortran
- There are various new options (e.g., a convergence window with a tolerance can be specified as a stopping condition, and the user can specify a subroutine for reporting the best solution at each generation).
- Mapping the variables to be between 0 and 1 now occurs internally, rather than requiring the user to do it.
- Can now include an initial guess in the initial population.
This is another example of how older Fortran codes can be refactored to bring them up to date with modern computing concepts without doing that much work. In my opinion, the object-oriented Fortran 2003 interface is a lot better than the old FORTRAN 77 version. For one, it’s easier to follow. Also, data can now easily be passed into the user fitness function by extending the pikaia class and putting the data there. No common blocks necessary!
- Original Pikaia code and documentation [High Altitude Observatory]