Algorithmically probable mutations reproduce aspects of evolution such as convergence rate, genetic memory, modularity, diversity explosions, and mass extinction.
Santiago Hernández-OrozcoHector ZenilNarsis Aftab KianiPublished in: CoRR (2017)
Keyphrases
- convergence rate
- evolutionary process
- premature convergence
- learning rate
- mutation operator
- convergence speed
- step size
- genetic algorithm
- global convergence
- gradient method
- aspect oriented programming
- numerical stability
- evolutionary algorithm
- primal dual
- fitness function
- differential evolution
- evolutionary computation
- faster convergence rate
- variable step size