2 Even Parity

Machine learning is typically concerned with inducing a model that can correctly classify input and output pairs. The set of boolean problems provides a similar framework for genetic programming. Langdon and Poli (2002) provide a discussion of boolean program spaces, which is also related to the field of evolvable hardware and circuit design [Koza et al., 1999].

Within the class of boolean problems, the Even-$n$-Parity problem requires the correct classification of bit strings of length $n$ having an even number of $1$'s. For example, Figure 2.2 shows the truth table for Even-3-Parity, where an even number of $1$'s is classified as ``True''. In this thesis, the case of $n=5$ will be considered.

Figure 2.2: Truth table for the Even-3-Parity problem.