Controlled Static Trading with GE

Abstract

Previous chapters in this book have examined the evolution of constants with the aim of identifying the most effective method of constant creation and adaptation. These experiments were conducted, for the most part, unencumbered by the added complexity of having to evolve the constants in line with function terminals. This enabled the development of a clear picture as to the behaviour and performance of different constant creation methods. In this chapter, GE is embedded in a more complex environment – that of financial markets. The aim here is to generate trading rules to optimally navigate a price series. However, as before, controlled and static experiments are initially conducted so as to determine the efficacy of the method adopted and to gain insight into the approach’s behaviour without the complexity of a shifting-fitness landscape.