Inverse Optimal Control with Linearly-Solvable MDPs.
Krishnamurthy DvijothamEmanuel TodorovPublished in: ICML (2010)
Keyphrases
- optimal control
- average cost
- reinforcement learning
- markov decision processes
- dynamic programming
- infinite horizon
- policy iteration
- risk sensitive
- finite horizon
- control problems
- special case
- feedback control
- state space
- optimal control problems
- optimal policy
- np hard
- markov decision process
- control strategy
- approximate dynamic programming
- class of nonlinear systems
- average reward
- brownian motion
- markov decision problems
- linear programming
- control law
- multistage
- markov chain
- linear quadratic