DeepRead
Cover of Approximate dynamic programming

Approximate dynamic programming

Warren B. Powell · 2011

Mathematics and physics › Numerical and matrix computation

#12,677 most recommended

About the book

The book covers sequential decision problems under uncertainty, where choices made now affect future states and rewards, and how to solve them when the state space is too large for exact dynamic programming. It is aimed at researchers and practitioners in operations research and related fields, with applications such as logistics, transportation, resource allocation, and energy. Its central claim is that many large sequential decision problems can be handled by approximating the value function or the decision policy, using simulation, statistical learning, and optimization together, instead of solving Bellman's equation exactly. It argues that the curses of dimensionality (large state, outcome, and action spaces) are the main obstacle and that a careful modeling framework, including the post-decision state, makes approximation tractable. It also argues that problem modeling and policy design should be treated as distinct steps, so practitioners can choose among several policy classes rather than assuming one method fits all.

Written by Claude from what it knows of this book.

What Hacker News says 50% recommend · 0% critical

00.51201120122013201420152011 peak · 1
This bookAverage book in Numerical and matrix computation

In the commenters’ own words

What Open Library readers say

No Open Library reader has rated it yet.

  • 1 want to read
  • 0 reading
  • 0 have read

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
2011
ISBN
9780470604458
Editions known
2 at Open Library
Open Library
OL16212035W

Dynamic programming

Explore in the mosaic

Read next: its neighbours on the mosaic