Summary

This concludes our journey into the unsettling world of evolutionary computing. There is a lot more to evolutionary computing than genetic algorithms such as artificial life, swarm intelligence, or differential evolution. Moreover, our description of genetic algorithm did not include genetic programming that applies genetic operators to trees and directed graphs of expressions.

This chapter dealt with a review of the different NP problems, the key components and genetic operators, an application of the fitness score to financial trading strategy, and the subtle variation in encoding predicates. The chapter concluded with an overview of the advantages and risks of genetic algorithms.

Genetic algorithms are an important element of a special class of reinforcement learning introduced in the Learning classifiers systems section of Chapter 15, Reinforcement learning. The next two chapters describe the most common reinforcement learning techniques starting with Bayesian inference algorithms...