Approximate Dynamic Programming for Constrained Linear Systems: A Piecewise Quadratic Approximation Approach.
Kanghui HeTon J. J. van den BoomBart De SchutterPublished in: CoRR (2022)
Keyphrases
- linear systems
- approximate dynamic programming
- linear program
- sufficient conditions
- dynamical systems
- dynamic programming
- reinforcement learning
- step size
- coefficient matrix
- average cost
- sparse linear systems
- markov decision processes
- policy iteration
- control policy
- approximation algorithms
- objective function
- state space
- primal dual
- np hard
- evolutionary algorithm
- machine learning
- real time