Global Convergence of Policy Gradient for Linear-Quadratic Mean-Field Control/Game in Continuous Time.
Weichen WangJiequn HanZhuoran YangZhaoran WangPublished in: ICML (2021)
Keyphrases
- optimal control
- linear quadratic
- policy gradient
- global convergence
- global optimum
- dynamic programming
- dynamical systems
- convergence speed
- control strategy
- convergence rate
- reinforcement learning
- closed loop
- optimization methods
- markov random field
- control system
- neural network
- em algorithm
- nash equilibrium
- markov chain
- vector valued
- objective function