Create a 2nd EA and perform parameter tuning

gianlucatruda / evocomp

Assignments for VU Masters course in Evolutionary Computing

GNU General Public License v3.0

0 stars 0 forks source link

Create a 2nd EA and perform parameter tuning #18

Closed gianlucatruda closed 4 years ago

gianlucatruda commented 4 years ago

[ ] Configure a new EA by extending and modifying deap_specialist (see issue #7 for ideas)
[ ] Perform parameter tuning
[ ] Build a class (based on BaseEAInstance that implements __init__ and evolve functions.

Serafim179 commented 4 years ago

Pull request #20 is up for both this and #17 issues. New EA is created from MyDemoEAInstance in EA_demo.py

Serafim179 commented 4 years ago

Gaussian mutation probability seems to be optimal around 0.3 or 0.4. It helps in combating the diversity deterioration problem and doesn't kill off the quality. Sigma doesn't seems to make much impact on fitness, only on diversity, so it was left as 1.