Last change
on this file since 11194 was
6152,
checked in by bfarka, 14 years ago
|
added ecj and custom statistics to communicate with the okb services #1441
|
File size:
860 bytes
|
Line | |
---|
1 | This package shows two examples of evolutionary search using ECJ's spatially-embedded models. Both of them use the optimization in the Rosenbrock problem domain.
|
---|
2 |
|
---|
3 | First, we show how to use a regular spatially-embedded evolutionary computation model (file: spatial.params). The population is embedded in a 1-D ring (the only spatial model available in ECJ v14), and a spatial version of tournament selection is used for breeding.
|
---|
4 |
|
---|
5 | Second, we show hot to use a spatially-embedded coevolutionary model (file: spatialcoevolve.params). Here, both populations are embedded in 1-D rings of the same size. Aside from the evolution in each population (which is set as in the previous case), we also employ a spatial coevolutionary evaluator: individuals are evaluated in combination with partners that reside at neighboring locations in the other populations.
|
---|
Note: See
TracBrowser
for help on using the repository browser.