amino acid S, i.e., ߨሺݔ, Sሻ as well as the similarity between the

sidue of a peptide x and the amino acid Y, i.e., ߨሺݔ, Yሻ. The

as calculated using the smaller similarity,

߰ሺܠ, ܚሻൌminሼߨሺݔ, Sሻ, ߨሺݔ, Y

max function was defined as below,

߰&ሺܠ, ܚሻൌmax

ሼߨሺݔ, ݎሻሽ

(8.14)

RPN chromosome expression for this max function was shown as

߰&ሺܠ, ܚሻൌቄෑሺݎሻቅ&

(8.15)

ose an example of the max function was (bReK)&. In this example,

were the second and the fifth residue of the peptide x while R and

he amino acids used in this rule for two residues, respectively.

hows how the fitness of this rule was calculated. Again, two

es were calculated at first. They were the similarity between the

esidue of a peptide x and the amino acid R, i.e., ߨሺݔ, Rሻ as well

milarity between the fifth residue of a peptide x and the amino acid

ሺݔ, Kሻ. The fitness was calculated using the greater one between

arities,

߰&ሺܠ, ܚሻൌ݉ܽݔሼߨሺݔ, Rሻ, ߨሺݔ, K

eal use of this min-max function was more complicated. For

suppose an RPN chromosome of a GP min-max rule was

(bReK)&)+. It was composed of three calculations involving two

The calculation of the fitness was shown below. The fitness of

subunit was shown in the Equation (8.12) and the fitness of the

ubunit was shown in the Equation (8.14). The final fitness of this

ted rule was shown below, where the $ key was used to denote a

both the min and the max operators,

ܚሻൌmin൛minሼߨሺݔ, Yሻ, ߨሺݔ, Sሽ, maxሼߨሺݔ, Rሻ, ߨሺݔ, Kሽൟ