On tight bounds for function approximation error in risk-sensitive reinforcement learning

DOI10.1016/j.sysconle.2021.104899zbMath1478.91064MaRDI QIDQ2243003

Prasenjit Karmakar, Shalabh Bhatnagar

Publication date: 10 November 2021

Published in: Systems \& Control Letters (Search for Journal in Brave)

Full work available at URL: https://doi.org/10.1016/j.sysconle.2021.104899

zbMATH Keywords

stochastic optimal control; stochastic systems; Perron-Frobenius eigenvalue; function approximation; eigenvalue perturbation; risk-sensitive reinforcement learning

Mathematics Subject Classification ID

91B06: Decision theory

68T05: Learning and adaptive systems in artificial intelligence

93E20: Optimal stochastic control

Cites Work

• Unnamed Item
• Risk-averse dynamic programming for Markov decision processes
• On comparison of the Perron-Frobenius eigenvalues of two ML-matrices
• Multiplicative ergodicity and large deviations for an irreducible Markov chain.
• Markov decision processes with average-value-at-risk criteria
• A quantitative comparison of risk measures
• Stochastic approximation with `controlled Markov' noise
• Simulation-based optimization of Markov reward processes
• Comparing the Spectral Radii of Two Nonnegative Matrices
• On the Maximal Eigenvector of a Positive Matrix
• Risk-Sensitive Optimal Control for Markov Decision Processes with Monotone Cost
• Q-Learning for Risk-Sensitive Control
• Inequalities
• A sensitivity formula for risk-sensitive cost and the actor-critic algorithm