Initialisation and Grammar Design in Grammar-Guided Evolutionary Computation

TL;DR

This paper compares the effects of grammar design and initialisation methods on different search algorithms, finding CFG-GP to be more robust than GE and random search across benchmark problems.

Contribution

It provides a comparative analysis of grammar design and initialisation impacts on GE, random search, and CFG-GP, highlighting CFG-GP's robustness and potential for future research.

Findings

CFG-GP is less sensitive to initialisation and grammar design.

Simple parameter tuning can improve CFG-GP performance.

CFG-GP outperforms GE and random search in robustness.

Abstract

Grammars provide a convenient and powerful mechanism to define the space of possible solutions for a range of problems. However, when used in grammatical evolution (GE), great care must be taken in the design of a grammar to ensure that the polymorphic nature of the genotype-to-phenotype mapping does not impede search. Additionally, recent work has highlighted the importance of the initialisation method on GE's performance. While recent work has shed light on the matters of initialisation and grammar design with respect to GE, their impact on other methods, such as random search and context-free grammar genetic programming (CFG-GP), is largely unknown. This paper examines GE, random search and CFG-GP under a range of benchmark problems using several different initialisation routines and grammar designs. The results suggest that CFG-GP is less sensitive to initialisation and grammar design than both GE and random search: we also demonstrate that observed cases of poor performance by CFG-GP are managed through simple adjustment of tuning parameters. We conclude that CFG-GP is a strong base from which to conduct grammar-guided evolutionary search, and that future work should focus on understanding the parameter space of CFG-GP for better application.