
Genetic Algorithms Digest   Wednesday, 19 September 1990   Volume 4 : Issue 21

 - Send submissions to GA-List@AIC.NRL.NAVY.MIL
 - Send administrative requests to GA-List-Request@AIC.NRL.NAVY.MIL

Today's Topics:

	**** SPECIAL ISSUE OF WORKSHOP ABSTRACTS (PART 3) ****

	- Message from moderator
	- An Investigation of Genetic Operators for Sequencing Problems
		Barry R. Fox
	- Lookahead Planning and Latent Learning in Classifier Systems.
		Rick L. Riolo
	- Parallel genetic algorithms and combinatorial optimization
		Heinz Muhlenbein

******************************************************************************

CALENDAR OF GA-RELATED ACTIVITIES: (with GA-List issue reference)

Conference on Simulation of Adaptive Behavior, Paris (v4n17)  Sep 24-28, 1990
Workshop Parallel Prob Solving from Nature, W Germany (v4n18) Oct 1-3,   1990
2nd Intl Conf on Tools for AI, Washington, DC (v4n6)          Nov 6-9,   1990
4th Intl. Conference on Genetic Algorithms (v4n17)            Jul 14-17, 1991

(Send announcements of other activities to GA-List@aic.nrl.navy.mil)

******************************************************************************

Date: Wed, 19 Sept 1990
From: Alan C. Schultz (GA-List Moderator)
Subject: Another Special issue

	This is the third issue comprised of abstracts from talks
	presented at the Foundations of Genetic Algorithms Workshop
	held July 15-18, 1990, at Indiana University.  The workshop
	was organized by Gregory Rawlins.

	This is the final installment of abstracts from the workshop.

	--Alan

--------------------------------

Date: Mon, 27 Aug 90 16:40 EST
From: "Barry R. Fox" <0003415313@mcimail.com>

  An Investigation of Genetic Operators for Sequencing Problems

  Producing offspring for sequencing problems requires different recombination
  operators than those used on traditional bit string problems.  Little work
  has been done to ehplain how reordering operators recognize and preserve
  good "building blocks" for sequences.  We describe a model for looking at
  sequences which reveals the predecessor/successor relationship between any
  two elements.  This model may be used to compare two sequences and recognize
  the relations they have in common and then generate offspring which preserve
  some or all of these relations.

		 Mary Beth McMahon and Barry Fox

-----------------------------------------

Date: Mon, 27 Aug 90 09:46:02 EDT
From: Rick_Riolo@um.cc.umich.edu

   Lookahead Planning and Latent Learning in Classifier Systems.

   Rick L. Riolo
   University of Michigan
   Ann Arbor, MI, 48109, USA
   rick_riolo@um.cc.umich.edu

   Classifier systems (CSs) have been used to simulate
    and describe the behavior of adaptive organisms, animats, and robots.
   However, CS implementations to date have
    all been reactive systems, which use simple S-R rules 
    and which base their learning algorithms on 
    trial-and-error reinforcement techniques similar
    to the Hullian Law of Effect.
   While these systems have exhibited interesting behavior
    and good adaptive capacity, they cannot do other
    types of learning which require having explicit internal models
    of the external world, e.g., using complex plans as humans do,
    or doing ``latent learning'' of the type observed in rats.
   This paper describes a CS that
    is able to learn and use internal models both
    to greatly decrease the time to learn general sequential decision
    tasks and to enable the system to exhibit latent learning.

   (This is the abstract from a paper that will appear in Procedings 
    of the Confernce on Simulating Animal Behavior (SAB 90), 
    Paris, Sept 1990; that paper includes some of the results 
    presented at the FOGA workshop.) 

-------------------------------------

Date: Thu, 30 Aug 90 18:17:37 EDT
From: Heinz.Muehlenbein@K.GP.CS.CMU.EDU

  Parallel genetic algorithms and combinatorial optimization
  Heinz Muhlenbein

  Parallel genetic algorithms (PGA) use two major modifications compared
  to the genetic algorithm. Firstly, selection for mating is distributed.
  Individuals live in a 2-D world. Selection of a mate is done by each
  individual independently in its neighborhood. Secondly, each individual
  may improve its fitness  during its lifetime by e.g. local hill
  climbing.

  The PGA is totally asynchronous, running with maximal efficiency on MIMD
  parallel computers. This has been achieved by the fact that the
  individuals are active and not acted on like in the genetic algorithm.

  We will show the power of the PGA with three combinatorial problems:
  the traveling salesman problem, the autocorrelation problem, and
  the graph partitioning problem.
  In all these examples, the PGA has found solutions of very large
  problems, which are comparable or even better than any other solution
  found by other heuristics.
  This is proof by experimental results.  In the talk we will also outline
  why and when the PGA is succesful. Firstly, the success is depending on
  the genetic representation of the combinatorial problem. Secondly, a
  suitable crossover operator and an efficient local hill climbing method
  is important.  Abstractly, a PGA is a parallel search with
  information exchange between the individuals. If we represent the
  combinatorial problem as a fitness landscape in a certain configuration
  space, we see, that a PGA tries to jump from two local minima to a
  third, still better local minima, by using the crossover operator.
  This jump is (probabilisticly) succesful, if the fitness landscape has a
  certain correlation.  We will show the correlation for the traveling
  salesman problem by a configuration space analysis. The PGA explores
  implicitly the above correlation.

  The PGA is a very simple algorithm and can be applied to many
  combinatorial problems with great success. Nevertheless, its analysis
  is very complex and has implications also for biology and
  physics.

--------------------------------
End of Genetic Algorithms Digest
********************************
