Before starting, I really appreciated the README.txt file since it allows me to have a clear overview of your idea before reading the code.
The idea of dividing the solution space in different 'Islands' is quite interesting in my opinion and I also thought in using this approach.
The code is clear and well-commented.
It is quite difficult to find something that could be improved in your idea since the results obtained are the best we could achieve.
Just one thing to say, I believe that the number of parts in which the solution space is divided can be crucial in obtaining good results. Reporting through plots the different performance accordingly to the lenght of sub-solutions could be useful, just becasue I think that in some situations aggregating local optimal solutions do not allow to reach the global optimal one.
To conclude, congratulations for the results. Hope to collaborate with you in the future.
Before starting, I really appreciated the README.txt file since it allows me to have a clear overview of your idea before reading the code. The idea of dividing the solution space in different 'Islands' is quite interesting in my opinion and I also thought in using this approach.
The code is clear and well-commented.
It is quite difficult to find something that could be improved in your idea since the results obtained are the best we could achieve. Just one thing to say, I believe that the number of parts in which the solution space is divided can be crucial in obtaining good results. Reporting through plots the different performance accordingly to the lenght of sub-solutions could be useful, just becasue I think that in some situations aggregating local optimal solutions do not allow to reach the global optimal one.
To conclude, congratulations for the results. Hope to collaborate with you in the future.