Table 1.
Parameters used for the development of evolutionary grammars using PonyGE2
| Parameter | Parameter setting |
|---|---|
| Algorithm | NSGAII [15] |
| Population size | 300 |
| Elite size | 30 |
| Generations | 1500 |
| Crossover | subtreea |
| Crossover probability | 0.9 |
| Mutation | subtreea |
| Mutation events | 1 |
| Selection proportion | 0.5 |
| Fitness 1 | F1 |
| Fitness 2 | Minimizing the number of nodes |
| Maximum derivation tree initialization depth | 10 |
| Maximum derivation tree depth | 15 |
| Initialization strategy | PI grow [21] |
a The crossover strategy is analogous to the one-point operator but by mixing tree structures. The mutation operator is applied only to the population resulting from the crossover