This page uses cascading style sheets! Unfortunately Your browser does not support that. Thus the page layout may be different from the intended one, but no information is lost.
Homepage of Dr. Andreas Frick
Because of the shutdown of the domain uni-karlsruhe.de due on the end of march 2014 also this page will be no longer available thereafter. The successor of the former Universität Karlsruhe (TH), the Karlsruhe Institute of Technology (KIT), no longer will provide personal private pages for non students or projects. I'm trying to move the content of the pages to my blogger account as soon as possible. Please excuse any related inconveniences.
Currently I'm still striving to propel my research in the area of Theory of Optimization at the the Universität Karlsruhe (TH), Germany.
TSPGA is a simple evolution program for the symmetric traveling salesman problem (TSP) based on the freely available PGAPack genetic algorithm library. It uses a randomized version of Lin's 2-opt improvement strategy for the genetic operators uniform crossover and mutation. It accepts data in the format used in the TSPlib sample problem collection. A more detailed description can be found in my short article. I have written this program mainly for two reasons: I wanted to investigate whether one can use elements of prevalent improvement strategies also in evolution programs, and I needed a test implementation of an evolution program for a hard problem as a starting point for my further studies to investigate what happens during evolution. The source code is available on request.
Since most of the research on evolutionary algorithms is focussed on experimental studies, I decided to focus on theoretical models. Already John Holland, one of the founders of genetic algorithms, tried to explain their working by the famous schema theorem. But although Holland has been inspired by natural evolution, it seems that he did not study the already existing literature on population genetics. In fact the model he presented in his doctoral thesis is the Wright-Fisher model of population genetics already established at that time. Unfortunately in his calculations Holland did not take the properties of the underlying Markov chain into account. Hence he claims to have found a relation between the expected population fractions of two successive generations, but which can be shown to be only a deterministic approximation of the selection process. This inaccuracy has been propagated in the literature, and still there is some discussion in the community about the relevance of that theorem. But Holland's idea of propagation of schemata respectively patterns during the evolution is alright, although it is mainly caused by genetic drift. It leads to a rapid assimilation of the individuals, which can be observed in experiments by using suitable indices for monitoring. A throughout discussion of these subjects can be found in my thesis.
Since the theory of evolutionary computation has many parallels to the theory of natural evolution, I had to read a lot of literature about population genetics and related topics. As a result I believe that a real understanding of evolution only can be achieved by studying the mathematical models behind it. Unfortunately most text books about evolution neglect this view totally. Evolutionary processes also can be observed in nature often directly. Finally, some side effects of the stochastic processes are apparent in nature. Consequently I believe that intelligent design is no reasonable explanation of the origin of the different species found on earth and should be abandoned totally.
I always have been interested in animals and especially their
behavior. An important question is how to train animals without using
violence. The problem first has been solved by Monty Roberts for horses. In
2002 I watched him taming several misbehaving horses on a show
in Karlsruhe, which I found very impressing. There are still
people not believing in his skills and regard the whole approach
to be a fluke, but I think that's only envy. Jan Fennell
developed a similar method to interact with dogs and to behave
as a pack leader. Since I have been running regularly with the Golden
Fleck of a friend, I could try out her hints
and can approve their efficacy. Unfortunately
Fleck passed away in
2012, may she rest in peace. Interestingly both approaches have been
developed by nonacademics but people with a lot of experience and a
certain cleverness. Cesar Millan seems to use the
same techniques but to be commercially more successful. A famous
German dog listener is Christiane Rohn. Another
important researcher in this area is Temple
But behavior also has evolved in almost the same manner as the species did. Thus social evolution is an important part of biology. I have been faced to this subject by a series of articles in the German newspaper Frankfurter Allgemeine Zeitung by Eckart Voland. The development of behavior can also be simulated on a computer using situations from game theory. Robert Axelrod first ran such experiments using the repeated prisoner's dilemma game. There he observed the superiority of the famous "tit for tat" strategy. During my time at the Institute of Applied Informatics and Formal Description Methods I supervised several seminar papers trying to extend his approach to simulate a dilemma market situation. Unfortunately the results are still unpublished.
During my studies I wrote a few reports, articles and theses. Most of them are available online. Here comes a summary in a BibTeX like formatting with links to the online versions.
In 2005 during the election campaign for the German Bundestag Paul Kirchhof presented his concept of a revised income tax and called for comments and suggestions. Hence I summed up my thoughts and opinions in a short outline (in German), which I sent to him.
Usually scholars get no instructions on writing papers, but it is important to follow a manual. Thus I collected a small list of suitable books in a BibTeX like formatting
I have collected some books and articles helping to survive in the struggles of campus life. One may also have a close look on the pages of THESIS, an association of doctoral candidates.
During the last years I collected some information about my family. My motivation was a review of the genealogy database software Gene, which I eventually installed on my old Macintosh Classic II. It is very user friendly and has some nice features, but unfortunately there is no version for OS X. Until now I collected data of about 150 persons I'm related to. My father's branch of my family tree seems to be quite old although I did not investigate that deeply until now. The earliest document about my ancestors I could get is dated around 1850. A nice heraldic figure is attributed to my father's family branch, but there seems to be no evidence of its grant. I'm afraid of that it is just a copy by a heraldic figure painter from the template on folio 208, volume 5 of the ancient German book Erneuert- und Vermehrtes Wappen-Buch by Johann Siebmacher. Other families of the same name not related to mine also keep a similar figure.
Aikido is a very effective art of self defense. It is rooted in the old Japanese martial arts but improves them by being totally nonviolent. I'm practicing Aikido for many years in a small group at the Post Südstadt Karlsruhe e. V. Since 1999 I'm the deputy coach there, and I started the first version of our WWW-pages. In 2000 I have been graduated to the 1. Kyu to wear a brown belt and later on in the same year I got a trainer license from the Deutscher Aikido Bund e. V. Currently I'm also managing the WWW-Pages, the calendar and the mail list of the group. Finally I have collected some hints to include the japanese kanji of the word aikido into WWW-pages (in German).
Schirmer/Mosel Verlag is the leading publisher of photographic art books in German.
Here are some hints to equipment I can recommend because of personal experience:
I'm a veteran Macintosh user, since I bought my Macintosh Classic in 1991 to write my first thesis and upgraded it to be a Macintosh Classic II in 1993. Although it is not very fast, it still runs much smoother than a comparable IBM compatible PC of the same age. I have used it mostly for typing texts at home and drawing figures with ClarisWorks. Recently I switched to a MacBook Pro, which a could get as a bargain after it has been discontinued. I'm very happy to have got it, not only because I saved a lot of money, but even a lot more because it is the last model without a glossy screen by default. I don't understand Apple's switch, because glossy screens generally are not suitable for work.
The following pages may be useful:
Feel free to send any hints and suggestions to my (old, but still working) e-mail address Andreas.Frick(at)wiwi.uni-karlsruhe.de or meet me on the following social networks:
If You like, You can share this page on these networks by clicking the buttons:
This page has been updated on Wednesday, 01-Jan-2014 21:42:27 CET.