Controlgym: Large-scale control environments for benchmarking reinforcement learning algorithms.
Xiangyuan ZhangWeichao MaoSaviz MowlaviMouhacine BenosmanTamer BasarPublished in: L4DC (2024)
Keyphrases
- reinforcement learning algorithms
- reinforcement learning
- model free
- markov decision processes
- real world
- state space
- dynamic environments
- control problems
- reinforcement learning problems
- control strategy
- multi agent environments
- reinforcement learning methods
- optimal control
- neural network
- temporal difference
- control system
- eligibility traces
- partially observable environments
- function approximation
- dynamic programming
- learning algorithm
- generative model
- linear programming
- control strategies
- objective function
- policy search
- training data