|
Last change
on this file since 1329 was
1306,
checked in by Maciej Komosinski, 8 months ago
|
|
Introduced symbolic names for dissimilarity estimation methods
|
|
File size:
1.5 KB
|
| Line | |
|---|
| 1 | rem To learn about all available options of each .py algorithm below, add "-h" to its parameters.
|
|---|
| 2 | rem Use the source code of the examples as a starting point for your customizations.
|
|---|
| 3 | rem Example usage:
|
|---|
| 4 |
|
|---|
| 5 | set DIR_WITH_FRAMS_LIBRARY=.................
|
|---|
| 6 |
|
|---|
| 7 |
|
|---|
| 8 |
|
|---|
| 9 |
|
|---|
| 10 | rem evolution with niching
|
|---|
| 11 | python -m evolalg.run_frams_niching -path %DIR_WITH_FRAMS_LIBRARY% -sim "eval-allcriteria.sim;deterministic.sim;sample-period-longest.sim" -opt velocity -genformat 0 -dissim PHENE_STRUCT_OPTIM -fit knn_niching -archive 50 -fitness_set_negative_to_zero -max_numparts 15 -max_numneurons 15 -max_numjoints 30 -max_numconnections 30 -max_numgenochars 10000 -popsize 50 -generations 10 -normalize none -hof_savefile HoF-niching.gen
|
|---|
| 12 |
|
|---|
| 13 | rem a generic island model example
|
|---|
| 14 | python -m evolalg.run_frams_islands -path %DIR_WITH_FRAMS_LIBRARY% -islands=10 -generations_migration=5 -sim "eval-allcriteria.sim;only-body.sim;deterministic.sim;sample-period-2.sim" -opt vertpos -genformat 1 -max_numparts 15 -max_numneurons 15 -max_numjoints 5 -max_numconnections 5 -max_numgenochars 100 -popsize 20 -generations 15 -hof_savefile HoF-islands.gen
|
|---|
| 15 |
|
|---|
| 16 | rem a combination of various parameters for thorough testing, see source of 'test_diferent_settings.py'
|
|---|
| 17 | rem python -m evolalg.tests.test_diferent_settings -path %DIR_WITH_FRAMS_LIBRARY% -sim "eval-allcriteria.sim"
|
|---|
| 18 |
|
|---|
| 19 |
|
|---|
| 20 |
|
|---|
| 21 | rem numerical examples (unrelated to Framsticks)
|
|---|
| 22 | python -m evolalg.run_numerical_example
|
|---|
| 23 | python -m evolalg.run_numerical_islands_example
|
|---|
Note: See
TracBrowser
for help on using the repository browser.
