 |
 |
Dear Sean, Thank you very much for your reply. It helped a great deal because it clarified more what I wanted to really do. What I really wanted to be able to do was evolve solutions without a GUI and then be able to view the performance of various solutions with a GUI. I did a little bit of hacking based on your suggestions and was able to create something along the lines of a SolutionViewer that extends my Problem but overrides setup() and evaluate() calling to the superclass versions when necessary. This has worked out very well although I'm quite sure it's a bit of a hack job. Unfortunately, while I'm an excellent java programmer...I have seen almost no Swing development time in the trenches. I think someday what I would like to see is something more integrated in terms of ECJ and MASON, a sort of ECWorkbench so that I can, for example, pick a population member while the evolutionary algorithm is being applied, and see what its performance was. Or see a graph of best and average fitness so far and select the best one for that generation and see its performance simulated. Or load an arbitrary individual and run it through the simulation especially if I could specify without parsing text files, that I want to see individual 22 of generation 34. As for ECJ, I'm curious if you ever end up with problems that run on the order of days? weeks? or are they just hours? minutes? Finally, does George Mason have a part-time Ph.D. program with courses at night? A link or other information?If so, do you have any comments about it? Thank you again for you time. Cordially, Steve