In collaboration with a professor of mine, we are starting to build a software toolbox providing implementations for stochastic optimization algorithms.

We would like to provide an open-source library/API in a way that would allow other programmers/researchers to experiment with different parameter tunings and also give them the ability to easily extend the available implementations with different techniques and other problems.

We focus mainly on Stochastic metaheuristic Algorithms such as: Simulated Annealing, Genetic Algorithms. Ant Colony Optimization and GRASP for tackling combinatorial optimization problems.

Our goal is to make this software as widely used as possible which means it has to be useful to its "users".

At this point it would be great to have some "user requirements" from the potential users - the OR community.

So, what would you like to have in a toolbox of this kind? What algorithms? What capabilities? What features would you make use of?

asked 09 May '12, 18:50

Florents%20Tselai's gravatar image

Florents Tselai
accept rate: 7%

edited 11 May '12, 12:20

Another addition to the increasing club of research groups that have invented their own stochastic optimization framework :) It's good to hear that you guys go open source: If you're ASL, BSD or MIT licensed, you can cross pollinate with the production focused frameworks (Drools Planner, jgap, ...) which have big, vibrant communities. It's surprisingly interesting (for both sides) to compare for example a ROADEF or ITC implementation.

(10 May '12, 10:36) Geoffrey De ... ♦

When I read the question, I thought of stochastic/simulation optimization problems, not stochastic algorithms. Not sure if I have ever heard of these algorithms being called "stochastic algorithms", but I understand why you may say so. My bad reading - sorry. Have removed my answer.

(13 May '12, 17:57) Samik R.

@Samik too bad you removed it, it was an interesting discussion and answer :)

(14 May '12, 04:41) Geoffrey De ... ♦

Since no one is answering this very interesting question with any new thoughts on user requirements, I might as well share my history on user requirements:

Here is the issue tracker with inputted user requirements of the users of Drools Planner (a friendly open source competing stochastic algorithm framework that is open to contributions :). Also, you 'll find more user requirements documented in the manual.


answered 11 May '12, 02:57

Geoffrey%20De%20Smet's gravatar image

Geoffrey De ... ♦
accept rate: 6%

edited 11 May '12, 03:25

@Geoffrey, how do you approach solving stochastic problems? Through simulation or through multi-stage recourse type formulations? Nice work on the solver, BTW - I need to take a closer look.

(11 May '12, 11:16) Samik R.

@samik Would you be so kind to clarify what exactly you mean with a stochastic problems (as opposed to a stochastic algorithms such as Simulated Annealing)? Preferably with some practical examples :) It could be that some of my users or even examples already use it for such use cases (or not at all), but I 'd like to get the terminology right before I can answer.

(11 May '12, 11:27) Geoffrey De ... ♦

@Geoffrey: Good point. See my comment to the original post.

(13 May '12, 17:58) Samik R.

@Samik You made me curious about what stochastic problems are, so I 've opened a new question for it.

(14 May '12, 04:45) Geoffrey De ... ♦
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: 09 May '12, 18:50

Seen: 2,250 times

Last updated: 14 May '12, 04:45

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