Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!clyde.concordia.ca!uunet!mcsun!hp4nl!dutrun!winfave
From: winfave@dutrun.UUCP (Alexander Verbraeck)
Newsgroups: comp.ai
Subject: Re: simmulated annealing
Keywords: annealing ai scheduling applications
Message-ID: <1031@dutrun.UUCP>
Date: 14 Dec 89 00:08:04 GMT
References: <558@granite3.UUCP>
Reply-To: winfave@dutrun.UUCP (A.Verbraeck)
Organization: Delft University of Technology, The Netherlands
Lines: 39

In article <558@granite3.UUCP> ahlenius@cell.mot.COM (Mark Ahlenius) writes:
>I am interested in real world applications for simulated
>annealing.  And was wondering if anyone has been using this
>technique in the "AI field".  If so what type of applications
>has it been used for.
>
>Specifically, has anyone used it for applications such
>as scheduling (like production, etc)?
>
>		Thanks,
>		'mark ahlenius

Yes, simulated annealing has been used in scheduling applications. I
have here in front of me the PhD-thesis of Peter van Laarhoven with the
title "Theoretical and Computational Aspects of Simulated Annealing". In
this thesis, an empirical analysis of the approach is used on the TSP
(Travelling Salesman Problem), the Job Shop Scheduling problem and the
Football Pool Problem. The conclusions of this chapter:

- the s.a. algorithm has a potential for finding high-quality solutions.
- it takes a lot of computing time
- s.a. is not a panacea: some sophisticated tailored algorithms from 
  literature are often competitive with and often superior to s.a.
- formulating a problem in a way that lends itself to application of
  simulated annealing is not a trivial task. Once an algorithm is
  found, the algorithm is easy to implement.

I myself haven't used the algorithm in solving production planning
(scheduling) problems yet. I plan to do so in the near future.

If anyone is interested in more information, post or e-mail. I can
perhaps dig up some more literature on this topic.

----------------------------------------------------------------------
Alexander Verbraeck                  e-mail: winfave@dutrun.tudelft.nl
Delft University of Technology               winfave@hdetud1.bitnet
Department of Information Systems            winfave@dutrun.uucp
PO Box 356, 2600 AJ  The Netherlands         dutrun!winfave@hp4nl.uucp
----------------------------------------------------------------------