propeller.problems.valiant

Possibly impossible to solve with genetic programming. Stems from the work of Leslie Valiant and involves determining the parity of an unknown subsequence of a larger sequence of bits.

-main

(-main & args)

Runs the top-level genetic programming function, giving it a map of arguments with defaults that can be overridden from the command line or through a passed map.

error-function

(error-function argmap data individual)

Finds the behaviors and errors of an individual: Error is 0 if the value and the program’s selected behavior match, or 1 if they differ. The behavior is here defined as the final top item on the BOOLEAN stack.

instructions

A list of instructions which includes keyword strings with the format “in + i” where i is a number from 0 to num-vars-1 concatenated with boolean and exec_if instructions and close.

train-and-test-data

Inputs are num-train random boolean values and outputs are the even parity of a subset of input variables.