Books, Publications, Reports and Presentations - Hartmut Pohlheim

  1. Ph.D. Thesis
  2. Genetic and Evolutionary Algorithm Toolbox for use with MATLAB
  3. Evolutionary Algorithms
  4. Genetic Programming
  5. Application of Optimization Techniques

Dissertation (Ph.D. Thesis)

* "Entwicklung und systemtechnische Anwendung Evolutionärer Algorithmen"
        (Development and Engineering Application of Evolutionary Algorithms)

Pohlheim, H.: Entwicklung und systemtechnische Anwendung Evolutionärer Algorithmen.
Aachen: Shaker Verlag, 1998. (ISBN 3-8265-4097-2)
zugleich: Dissertation, Technische Universität Ilmenau, 1998. (written in german)
Das Buch kann direkt beim Shaker Verlag online bestellt werden oder über jeden Buchladen (The book can be ordered directly from Shaker Verlag or by every bookstore):
Hartmut Pohlheim, Entwicklung und systemtechnische Anwendung Evolutionärer Algorithmen, ISBN 3-8265-4097-2; Preis: 98,-DM.
Für weitergehende Informationen steht eine spezielle Dissertations-Seite zur Verfügung. Diese umfaßt Verweise zu Inhaltsverzeichnis, umfassenden Leseproben, einer Zusammenfassung (Thesen) in deutsch und englisch sowie Angaben zur Bestellung der Arbeit über den Buchhandel. (For more information about my Ph.D. thesis please consult the special thesis page. This page contains links to the table of contents of the thesis, large excerpts of the thesis, a summary of the thesis in english and directions for ordering the book online from Shaker Verlag.)

Genetic and Evolutionary Algorithm Toolbox for use with MATLAB

* Genetic and Evolutionary Algorithm Toolbox (GEATbx): Documentation

Pohlheim, H.: Genetic and Evolutionary Algorithm Toolbox for use with Matlab. Technical Report, Technical University Ilmenau, 1994-1998.
english - frequently updated -
HTML format
The whole documentation consists of a fact sheet, a Tutorial, an Introduction to Evolutionary Algorithms, Examples of objective functions and a complete reference of the GEA Toolbox including Reference of GEA Toolbox functions, options settings and the used references (literature).
The Genetic and Evolutionary Algorithm Toolbox is not public domain. If you are interested in using or testing the toolbox please contact the author

* Genetic Algorithm Toolbox, earlier papers and reports

Chipperfield, A., Fleming, P. J., Pohlheim, H. and Fonseca, C. M.: Genetic Algorithm Toolbox for use with Matlab. Technical Report No. 512, Department of Automatic Control and Systems Engineering, University of Sheffield, 1994.
english - compressed Postscript

Chipperfield, A. J., Fleming, P. J. and Pohlheim, H.: A Genetic Algorithm Toolbox for MATLAB. Proc. Int. Conf. Sys. Engineering, Coventry, UK, 6-8 Sept., pp. 200-207, 1994.

Evolutionary Algorithms

* Evolutionary Algorithms: Principles, Methods and Algorithms

Pohlheim, H.: Evolutionary Algorithms: Principles, Methods and Algorithms. Technical Report, Technical University Ilmenau, 1994-1997.
english - frequently updated - HTML format
Abstract: An introduction into Evolutionary Algorithms: structure, operators and research results; can be used as introduction or as reference of algorithms; uses many graphics.


* Publications on Theory and Application of Evolutionary Algorithms

Pohlheim, H.: Advanced Techniques for the Visualization of Evolutionary Algorithms Proceedings of 42. International Scientific Colloquium Ilmenau, vol. 3, pp. 60-68, 1997.
english - September 1997 - Postscript - compressed Postscript
Abstract: Evolutionary algorithms work in an algorithmically simple manner. When put to work they produce a vast amount of data. Apart from simple convergence information it is a non-trivial task to extract useful information from those data to provide insight into the state and progress of the evolutionary algorithm. Methods for extracting and visualizing relevant data are still under development.
In this paper an advanced technique for visualizing multidimensional data produced by the evolutionary algorithm is presented. This advanced technique is independent of the representation of the variables. The technique is applied to the visualization of variables of individuals of a population. The use of this technique for the visualization of multi criteria objective values in multiobjective optimization is demonstrated. Especially useful is this technique for comparison of high-dimensional data of different runs.
Beside the advanced method a set of standard visualization techniques for different data types is presented. These data types and techniques give an instant visual impression of the evolutionary algorithms progress and the actual state of the individuals of the population. The methods were selected according to their usefulness for real world applications. Hopefully, the defined standard set will open up a discussion between designers and users of evolutionary algorithms about necessary and useful visualization techniques.

Slides of the presentation - english - September 1997 - PowerPoint Animation - Postscript - compressed Postscript

Pohlheim, H.: Ein genetischer Algorithmus mit Mehrfachpopulationen zur Numerischen Optimierung. at-Automatisierungstechnik 3 (1995), pp. 127-135, 1995.
german - September 1994 - Postscript - compressed Postscript
Abstract: A genetic algorithm with multiple population (MPGA) is described. The operators of the algorithms are discussed in detail. Especially the multiple population concept is explained.
A new method for fitness assignment by ranking using a non-linear distribution is introduced. The use of non-linear ranking permits higher selective pressures than the conventional ranking methods.
The use of the genetic algorithm in dynamic and control system optimization is presented together with an example.
The algorithm was implemented in MATLAB and is available as part of the Genetic and Evolutionary Algorithm Toolbox.


*Technical Reports on Theory and Application of Evolutionary Algorithms

Pohlheim, H.: Control of Lateral Vehicle Dynamics and Dynamic Optimization using the Genetic Algorithm Toolbox
Technical Report - english - March 1995 - HTML format
Abstract: The first part gives an overcview of the Genetic Algorithm Toolbox. Using two examples the use of the Toolbox is demonstrated for different problem domains: lateral control of a vehicle for parameter optimization and the push-cart system for dynamic optimization.

Pohlheim, H.: The Multipopulation Genetic Algorithm: Local selection and Migration
Technical Report - english - March 1995 - HTML format
Abstract: The papers gives an overview of the Multipopulation Genetic Algorithm followed by an detailed discussion of non-linear ranking, local selection und migration.


*Presentations on Evolutionary Algorithms

Pohlheim, H.: Evolutionäre Algorithmen - Optimierung nach Prinzipien der Natur (Evolutionary Algorithms - Optimization by natural principles)
Presentation - german - June 2000 - PowerPoint Animation - Postscript - compressed Postscript
Vorlesung an der Technischen Universität Ilmenau für Studenten im Hauptstudium Automatisierungs- und Systemtechnik (lecture at the Technical University Ilmenau).

Pohlheim, H.: Evolutionäre Algorithmen - Funktion, Struktur, Verfahren
Presentation - german - March 1995 - Postscript - compressed Postscript
Vortrag über Evolutionäre Algorithmen, historische Entwicklung, Struktur, Bestandteile, Funktionsweise der verschiedenen Operatoren sowie eine kurze Übersicht zur Genetic Algorithm Toolbox.

Genetic Programming

*Generation of structured process models using Genetic Programming

Pohlheim, H. and Marenbach, P.: Generation of structured process models using Genetic Programming. in Proceedings of AISB Workshop on Evolutionary Computing 1996, volume 1143 of Lecture Notes in Computer Science, Berlin, Heidelberg, New York: Springer-Verlag,, pp. 102-109, 1996.
english - February 1996 - Postscript - compressed Postscript
Abstract: The design of structured mathematical models of processes in a certain level of abstraction defined by the given task appears to be difficult and time consuming even for experienced experts.
This paper reports on a new method for the design of structured process models based on the metaphor of Genetic Programming. This new methodology allows the automatic generation of non-linear process models in a self-organizing way.

Slides of the presentation - english - April 1996 - PowerPoint Animation - Postscript - compressed Postscript

Application of Optimization Techniques

* Optimal control of greenhouse climate using Evolutionary Algorithms

Pohlheim, H. and Heißner, A.: Optimal Control of Greenhouse Climate using a Short Time Greenhouse Climate Model and Evolutionary Algorithms. in Proceedings of 3rd IFAC/ISHS Workshop on "Mathematical and Control Applications in Agriculture and Horticulture" Hannover, pp. 113-118, 1997.
english - October 1997 - paper not freely available (copyright IFAC),
please contact the author or have a look at the next paper

Pohlheim, H. and Heißner, A.: Optimal Control of Greenhouse Climate using Evolutionary Algorithms. in Proceedings of 42. International Scientific Colloquium Ilmenau, vol. 3, pp. 9-16, 1997.
english - September 1997 - Postscript - compressed Postscript
Abstract: The use of evolutionary algorithms for calculation of the optimal control of the states of a greenhouse system will be presented. The integrated model employed (greenhouse climate, crop growth, outside weather conditions and control equipment) predicts temperature, air humidity and CO2 concentration in a time interval of 15-60 minutes (short time-scale model). The paper presents the optimization of the control of the greenhouse climate to maximize the profit under certain constraints (for instance, prevention of stress for the crops) using evolutionary algorithms. By incorporation of problem specific knowledge into the evolutionary algorithm better results were produced in a shorter time. The results of optimization for optimal control using real world weather data are shown.
Slides of the presentation - english - September 1997 - PowerPoint Animation - Postscript - compressed Postscript

Pohlheim, H. and Heißner, A.: Anwendung Genetischer Algorithmen zur optimalen Steuerung des Gewächshauses. in GMA-Kongreß'96, VDI-Berichte 1282, pp. 799-809, Düsseldorf: VDI-Verlag, 1996.
german - July 1996 - Postscript - compressed Postscript
Abstract: Die Verwendung Genetischer Algorithmen für die Berechnung der optimalen Steuerung eines komplexen dynamischen Systems wird am Beispiel der Steuerung der Zustandsgrößen in einem Gewächshaus dargestellt. Grundlage ist ein integriertes Modell (Gewächshausklima, Wachstum und Transpiration des Pflanzenbestandes) für die Vorhersage von Temperatur, Luftfeuchtigkeit und Kohlendioxidkonzentration im Gewächshaus in einem Zeitbereich von 15-60 Minuten (Kurzfristmodell).
Die Arbeit zeigt die Optimierung der Steuerung des Gewächshausklimas mit dem Ziel einer Maximierung des Gewinns unter Berücksichtigung von Beschränkungen (z.B. Verhinderung von Streß für die Pflanzen). Durch die Einbeziehung von aufgabenspezifischem Wissen in den Genetischen Algorithmus gelang es, wesentlich schneller und zu besseren Ergebnissen zu gelangen, als dies mit einem normalen genetischen Algorithmus möglich war. Es werden Ergebnisse der Optimierung für durchschnittliche Tage verschiedener Jahreszeiten sowie Optimierungsergebnisse bei sich ändernden Preisen gezeigt.

Slides of the short presentation - german - September 1996 - PowerPoint Animation - Postscript - compressed Postscript
presented Poster - german - September 1996 - PowerPoint Animation - Postscript - compressed Postscript

Pohlheim, H. and Heißner, A.: Optimal Control of Greenhouse Climate using Genetic Algorithms. in MENDEL'96 - 2nd International Conference on Genetic Algorithms. 26.-28. June 1996, Technical University of Brno, Czech Republik,, pp. 112-119, 1996.
english - April 1996 - Postscript - compressed Postscript
Abstract: To understand greenhouse systems complex models are used increasingly. This paper presents an integrated model incorporating all components of a greenhouse: greenhouse climate, crop growth, outside weather conditions and control equipment. The model predicts temperature, air humidity and CO2 concentration in a time interval of 15-60 minutes. This model can be used for short time tasks as maximization of growth or prevention of stress. Using genetic algorithms the control is optimized for maximal profit. The results of optimization for optimal control during three average days and different control strategies by changing costs of control are shown.
Slides of the presentation - english - June 1996 - PowerPoint Animation - Postscript - compressed Postscript


*Simulation und Optimierung eines Blaualgen-Wachstums-Modells (blue-algae growth model)

Pohlheim, H.: Simulation und Optimierung eines Blaualgen-Wachstums-Modells. Diplomarbeit, Technische Universität Ilmenau, 1993.
(Simulation and Optimization of a blue-algae growth model. diploma, in german)

© 1998-2001 Hartmut Pohlheim, hartmut@pohlheim.com, www.pohlheim.com (last update: 03/2001)