DeepRead
Cover of Controlled Markov processes

Controlled Markov processes

E. B. Dynkin · 1979

Machine learning, statistics and probability › Bayesian and probabilistic modeling

#6,649 most recommended

About the book

The subject is the optimal control of Markov processes, in which a decision-maker picks actions that influence the transition law of a random process and aims to optimise an expected reward or cost. The scope covers discrete-time Markov decision models and controlled processes on general measurable state spaces, together with the probability and measure-theoretic foundations they need. The book is co-authored with A. A. Yushkevich. The book treats control as a problem over strategies and builds the theory around the dynamic programming principle and the functional equations it yields. Its central claim is that, under suitable conditions, the value of a control problem satisfies such equations and that optimal or near-optimal policies can be found among simple Markov policies. It examines criteria such as total, discounted and average reward, and sets out the regularity conditions (semicontinuity, compactness of action sets, measurable selection) needed to justify these results.

Written by Claude from what it knows of this book.

What Hacker News says 25% recommend · 0% critical

01220102015202020252011 peak · 2
This bookAverage book in Bayesian and probabilistic modeling

In the commenters’ own words

What Open Library readers say

No Open Library reader has rated it yet.

Open Library’s page for the book

What critics say, via Wikipedia

No Wikipedia article about this book yet, so no critics to quote.

Reddit communities — coming later

Details

First published
1979
ISBN
9780387903873
Editions known
2 at Open Library
Open Library
OL16348307W

Control theoryMarkov processesMathematical optimization

Explore in the mosaic

Read next: its neighbours on the mosaic