In Genetic Programming $($GP$),$ you create software that can synthesize new programs

expressed as tree structures to serve specific functionality $($purpose$).$ You provide this

functionality through an objective function for a specific program $($being synthesized$)$ to

your GP software by

a$)$ providing the closed form algebraic expression that your new program is

supposed to codify

b$)$ providing many training examples of input and output data that manifest the

desired functionality

c$)$ providing a series of constraint relationships between variables that should be

captured in the new program

d$)$ severely and correctly limit your terminal set $($leaves$)$ and function set $($nodes$)$ so

that the program automatically synthesizes to the desired one.