Information Technology Reference
In-Depth Information
As you can see, the GEP-PO algorithm can also jump from peak to peak,
finding along the way higher and higher positions in the solution landscape.
For instance, the thirteen best solutions designed with the GEP-PO algo-
rithm are better than the best solution discovered with the HZero algorithm,
f
(-5.457550049, -5.648040772, 0.579986572, 0.549194336, 9.863006592)
= 2.74912317067176, which again emphasizes that whenever great preci-
sion is needed the more computationally demanding GEP-PO algorithm
should be used.
In the next chapter we will study yet another interesting application of
gene expression programming - decision tree induction - where random
numerical constants also play an important role.
