Jump to content

Talk:Stochastic programming

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Description

[edit]

I do not agree with this description of stochastic programming. What is decribed here is rather "randomized methods", or simulation-type methods with stochastic elements. Stochastic programming is an area in optimization, or ,more specifically, in mathematical programming.

In fact Monte-Carlo, Simulated Annealing, Genetic Algorithm, ... are _stochastic_algorithms_ but not always used for "stochastic programming". A mathematical program is a model which leads to a solution by the use of an algorithm. So speaking, "stochastic programming" is a stochastic model which is used to feed an algorithm.

Let's define :

"Stochastic Algorithms"

"Stochastic Programming"

Lannez 11:03, 11 April 2007 (UTC)[reply]

Title?

[edit]

Maybe this page (which is currently just a stub) should instead be titled Stochastic dynamic programming, and should refer to dynamic programming methods applied to problems that also involve random shocks? --Rinconsoleao (talk) 10:02, 20 May 2009 (UTC)[reply]

I'm fine with a page move/change of name. My experience with SDP models in biology doesn't square with much of the lede (for example "Whereas deterministic optimization problems are formulated with known parameters, real world problems almost invariably include some unknown parameters" implies to me that the optimizing agent does not know the probability with which the stochastic events happen), and "random shocks" isn't a term that gets applied where I come from. I'm happy to leave the bulk of this article to the Econ crowd to reflect the topic from that tradition. Pete.Hurd (talk) 20:19, 20 May 2009 (UTC)[reply]

Textbooks

[edit]

I added some standard textbooks on "stochastic programming". It would be useful for others to help by using the "cite book" ( {{cite book}} ) template and adding links to Mathematical Reviews. Prekopa's name needs an accent, etc.

Please consider whether all of the earlier references are needed. It is not clear to me that unpublished working papers meet the criteria for inclusion, when this field is so established.

Thank you for your consideration. Best regards, Kiefer.Wolfowitz (talk) 04:16, 22 September 2010 (UTC)[reply]

Markov Decision Process

[edit]

These are the same thing has Markov Decision Processes. Or Semi-Markov Decision Processes if you use Continuous Time Stochastic Programming instead of discrete states. 129.10.244.231 (talk) 17:02, 29 April 2011 (UTC)[reply]

Economic applications

[edit]

I think that one important application to be mentioned is the optimal stopping problem and the associated Nested Fixed Point Algorithm as developed by Rust (1987; Econometrica). An important reference on economic applications is Adda and Cooper's book from 2003 (ISBN-13: 978-0262012010). Superpronker (talk) 19:22, 10 April 2012 (UTC)[reply]

Scope of the article

[edit]

Currently, this article focuses very heavily on stochastic linear programming (plus a very detailed and technical section on statistical inference). It does not any of the following, which may or may not be relevant on this page:

Some of these already have articles. Which topics should this article (not) address?

Dmildy (talk) 00:01, 23 November 2020 (UTC)[reply]