Evolving and Analyzing Modularity with GLEAM (Genetic Learning by Extraction and Absorption of Modules)

General methods for the evolution of programs with modular structure have long been sought by genetic programming researchers, in part because modularity has long been considered to be essential, or at least helpful, for human programmers when they develop large-scale software projects. Multiple efforts have been made in this direction, and while success has been demonstrated in specific contexts, no general scheme has yet been demonstrated to provide benefits for evolutionary program synthesis that are similar in generality and significance to those provided by modularity in human software engineering. In this chapter, we present and analyze a new framework for the study of the evolution of modularity, called GLEAM (Genetic Learning by Extraction and Absorption of Modules). GLEAM’s flexible architecture and tunable parameters allow researchers to test different methods related to the generation, propagation, and use of modules in genetic programming.