A Gene Expression Programming Implementation in Python
Mk II v1 is the only completed version
To use, give a list of lengths of the expression genes sequences, a list of lengths of the homeotic genes sequences, and a list of one or two genomes, which will be used either for both groups of genes sequences, or the first for the set of expression sequences and the second for the homeotic sequences.
To construct a genome, instatiate the class and then add a series of characters, functions, and arity, or a single terminal. The function should accept a list of inputs that are all numerical values, and use them to compute. The ordering of the inputs is based on the order which the node recieved the value, usually from its leftmost child node to its rightmost child node, if the tree were to be drawn out.
To run, give a fitness function, which will be used to compute the fitness of each organism, in the form of a chromosome, each generation. It should have all of the inputs and desired outputs allready builtin, or if being used in an unsupervised setting, be able to access the members of the organism that it requires to properly rate it. It will be called every round, being passed an organism, and if the organism requires other members to be tested properly, it should be of a subclass of the Chromosome class. It can also be passed another value whose default value should be False, and when it is True, the environment is asking for the maximum fitness value an organism can have. If in an unsupervised setting, then the function should return a value nearing infinity or that is unreachable, otherwise the simulation may preemptively stop. In unsupervised cases, the fitness function should have the power to stop the program from running by calling exit(), based on any parameter that it has access to.