Context Matters: Adaptive Mutation for Grammars

This work proposes Adaptive Facilitated Mutation, a self-adaptive mutation method for Structured Grammatical Evolution (SGE), biologically inspired by the theory of facilitated variation. In SGE, genotype individuals contains list each non-terminal grammar that defines search space. our proposed mutation, individual an array with different, rate non-terminal. We also propose Function Grouped Grammars, design procedure, to enhance benefits mutation. Experiments were conducted on three symbolic regression benchmarks using Probabilistic (PSGE), variant SGE. Results show approach is similar or better when compared standard and