Excellent Idea! This is exactly what is needed to solve problems which is hanging around us for quite some time now....
In my views we should devise new genetic operators and try them in our experiments... since we already have a large base of experiments running... also it would be easier to compare there performances...
Another point is also that the design of these new Pack of Genetic Operators should be such that they can be plugged into all our existing project as well as new without much re-code.
Remember your post about Alleles.. I have been wondering about that for quite some time but never came out with something workable... I think we may develop that too....
Another idea which struck me long back and I had plans to use it in my next project was after crossover or mutation - calculating a measure of diversity between parent chrome and the resulting child... and that way retaining only diverse population.... Considering a solution is always a N-dimensional hyper-space... this may easily be calculated using Euclidean Distance between two N-dimansional points.....
Mitchell I hope to release a small program which does SCRA analysis soon... this will tell us the number of FCBA, or pruned or perfect network... will draw out various statistics as to howmuch of weight matrix is saved... all with respect to number of Nodes in SCRA and Skip and Advance parameters...
--
------------------------------------------------------------------------- Using Tomcat but need to do more? Need to support web services, security? Get stuff done quickly with pre-integrated technology to make your job easier Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
