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

DOI10.1016/J.SYSCONLE.2021.104899zbMATH Open1478.91064OpenAlexW3149920308MaRDI QIDQ2243003FDOQ2243003

Authors: 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

Recommendations

• A learning algorithm for risk-sensitive cost
• Advances in Artificial Intelligence
• Risk-sensitive reinforcement learning
• Performance Loss Bounds for Approximate Value Iteration with State Aggregation
• Error bounds for approximations from projected linear equations

zbMATH Keywords

stochastic optimal controlstochastic systemsfunction approximationeigenvalue perturbationPerron-Frobenius eigenvaluerisk-sensitive reinforcement learning

Mathematics Subject Classification ID

Learning and adaptive systems in artificial intelligence (68T05) Decision theory (91B06) Optimal stochastic control (93E20)

Cites Work

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

Cited In (4)

• Estimation and approximation bounds for gradient-based reinforcement learning
• On the sample complexity of actor-critic method for reinforcement learning with function approximation
• Advances in Artificial Intelligence
• Data-driven direct adaptive risk-sensitive control of stochastic systems

This page was built for publication: On tight bounds for function approximation error in risk-sensitive reinforcement learning

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2243003)