I'm designing a GP problem that has "ephemeral random constants" between
its terminals. An ephemeral random constant gets a random value when
created which is maintained constant during the rest of the evolution.
Can I implement this with ECJ? How?. I'm not sure if each node in each
GPTree is an instance in its own (which could then store values
different to other instances) or pointers to a common class (in that
case, how could I differentiate each node?).
I have a similar problem also with the terminals, in the examples a
different class is used for X and another for Y, in a two variable
regression problem. I have vectors of data of size 64 (V), and I want to
use a single class (namely, V[i]) and decide which value i has at
Any help would be greatly appreciated,