Policy Optimization in a Noisy Neighborhood: On Return Landscapes in Continuous Control.
Nate RahnPierluca D'OroHarley WiltzerPierre-Luc BaconMarc G. BellemarePublished in: NeurIPS (2023)
Keyphrases
- global optimization
- control policy
- optimal policy
- neural network
- control system
- control policies
- optimization method
- combinatorial optimization
- action selection
- optimization process
- action space
- control theory
- adaptive control
- constrained optimization
- control method
- optimization methods
- optimization algorithm
- optimization problems
- dynamic programming
- noisy environments
- control problems
- mobile robot
- discrete optimization
- direct search