I am looking for any worked examples on multi-stage stochastic programming with recourse but failed to find any. People most of the times assume a distribution is known (which in practice is not and a model of the evolution needs to be determined) and then describe all the rest. I know this is problem-specific but I would like to see examples of this from scratch, e.g. we have this raw historical data, we obtain this distribution, formulate the problem and solve it.

I might probably be asking for a lot but any resources would help.

asked 25 May '16, 18:05

lstavr's gravatar image

accept rate: 0%

The example I use for two-stage SLP is the farmer problem from Birge and Louveaux. Here is my favorite explanation from the top Google results: http://myweb.clemson.edu/~pbelott/bulk/teaching/lehigh/ie426-f09/lecture22.pdf

I am unaware of a complete integrative analytics treatment, i.e., one that looks at data analysis, forecasting, optimizing, and doing it again until it makes sense. But I like the idea so much, I might just write it myself!

You'll just have to wait a bit for that part of the answer...


answered 25 May '16, 21:29

Leo's gravatar image

accept rate: 8%

edited 25 May '16, 21:30

Thank you for this. My problem is that all examples start with: "Assume the probabilities are x" but I have not found working examples on how these probabilities are derived in the first place (e.g. in my case, I have a number of discrete actions to select for which I have historical daily data on the function I want to optimize and a function which is the LHS of constraints from time series -- both as functional time-series -- and I want to derive the scenarios and run an optimization for the future). If you write something, please let me know -)

(26 May '16, 07:31) lstavr

I think what you need is a research paper or thesis with a case study. A thesis with a case study should have all the components you need. However papers are different as they usually focus on a particular aspect of stochastic programming (e.g., modeling, solution, and scenario generation) and not all the aspects. Assuming you are not that interested in solution methods, papers focusing on scenario generation for a specific applied problem could be of interest. A relatively complete example is available from here. It doesn't have all the details, but it could be a good start.


answered 26 May '16, 01:50

Ehsan's gravatar image

Ehsan ♦
accept rate: 16%

Thank you a lot, this is true. I will have a look at the paper!

(26 May '16, 07:32) lstavr
Your answer
toggle preview

Follow this question

By Email:

Once you sign in you will be able to subscribe for any updates here



Answers and Comments

Markdown Basics

  • *italic* or _italic_
  • **bold** or __bold__
  • link:[text](http://url.com/ "Title")
  • image?![alt text](/path/img.jpg "Title")
  • numbered list: 1. Foo 2. Bar
  • to add a line break simply add two spaces to where you would like the new line to be.
  • basic HTML tags are also supported



Asked: 25 May '16, 18:05

Seen: 1,282 times

Last updated: 26 May '16, 07:32

OR-Exchange! Your site for questions, answers, and announcements about operations research.