Step size adaptation in evolution strategies using reinforcement learning.
Sibylle D. MüllerNicol N. SchraudolphPetros KoumoutsakosPublished in: IEEE Congress on Evolutionary Computation (2002)
Keyphrases
- step size
- evolution strategy
- temporal difference
- reinforcement learning
- evolutionary programming
- convergence speed
- differential evolution
- evolutionary algorithm
- global search
- convergence rate
- particle swarm optimization algorithm
- cost function
- adaptation process
- numerical optimization
- genetic algorithm
- faster convergence
- function approximation
- stochastic gradient descent
- state space
- optimization methods
- mutation operator
- gradient method
- variable step size
- particle swarm optimization
- machine learning
- particle swarm optimization pso
- learning algorithm
- image compression
- linear programming
- model free
- optimization problems
- dynamic programming
- multi objective
- similarity measure
- image segmentation
- line search
- neural network