Dear all ECJ developpers,

I would like to implement a Local Search (LS loop) in ECJ during an EA 
loop so as to obtain a Memetic Algorithm.

The LS loop, which is able to generate and evaluate many individuals, 
could be applied after initialization and/or crossover and/or mutation 
with a certain probability.

According to you, what is the correct manner to implement the LS loop in 
ECJ ? Create a class which herits from "Breeder" ?

I think it is completely different of new "singlestate" feature in ECJ v25.

Thank you in advance for all information that can help me.



             (' O-O ')
Denis PALLEZ           [log in to unmask]
Maitre de Conférences / Associate Professor
Université Côte d’Azur (IUT), CNRS, I3S
(+33) 4 92 96 51 27