A performance gradient perspective on gradient‐based policy iteration and a modified value iteration

DOI10.1108/17563780810919096MaRDI QIDQ3613729zbMATH OpenOpenAlexFDO

Authors Lei Yang, James Dankert, Jennie Si

Publication date 12 March 2009

Published in International Journal of Intelligent Computing and Cybernetics (Search for Journal in Brave)

Full work available at URL https://doi.org/10.1108/17563780810919096

zbMATH Keywords

gradient methodsiterative methodsMarkov processesprogramming and algorithm theory

Mathematics Subject Classification ID

Dynamic programming (90C39)

Recommendations

• scientific article; zbMATH DE number 1753152
• Stability and monotone convergence of generalised policy iteration for discrete-time linear quadratic regulations
• scientific article; zbMATH DE number 1753153
• Approximate gradient methods in policy-space optimization of Markov reward processes
• Adaptive policy-iteration and policy-value-iteration for discounted Markov decision processes

Cites work

• Planning and acting in partially observable stochastic domains
• Potential-Based Online Policy Iteration Algorithms for Markov Decision Processes

Cited in

(2)

• A basic formula for performance gradient estimation of semi-Markov decision processes
• Mean-payoff optimization in continuous-time Markov chains with parametric alarms

This page was built for publication: A performance gradient perspective on gradient‐based policy iteration and a modified value iteration

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3613729)