Decoration
max planck institut
informatik
mpii logo Minerva of the Max Planck Society

Benjamin Doerr: Teaching


Evolutionary Algorithms (winter 06/07)

PD Dr. Benjamin Doerr, Dr. Nils Hebbinghaus and Dr. Frank Neumann

Time: Tuesday 11-13. Start: October 17.

Room: E1.4 R.023

Content: Evolutionary algorithms, or more generally randomized search heuristics, have widely and successfully been applied to complex engineering problems and problems from combinatorial optimization.

In the lecture we introduce different randomized search heuristics that use concepts from nature. Besided evolutionary algorithms, these include simulated annealing or the metropolis algorithm. We analyze these algorithms from a theoretical point of view. Such analyses help to understand how these heuristics work as well as to design better such algorithms for new problems. This is a particularly interesting area since still most of the existing, very promising experimental results in this field need a theoretical justification/explanation.

Audience: The course is a 2 hours computer science lecture. It is a "Spezialvorlesung". Together with the exercises (also 2 hours), it yields 6LP (provided you pass the exam). The course needs no particular prerequisites except basics in algorithmics and mathematics. The lecture will be given in English.

Exercises: The exercise course takes place Thursday, 9:30, in Room 023 (starts October 26). You may form teams of two. Each exercise sheet has four problems. Fully solving a problem earns you four point. Not even trying to solve a problem earns you one point. To be admitted to the final exam, you need 62,5% of the points (in average 10 per sheet equivalent to fully solving two problems). Exercise sheets will be posted below on Wednesday. They are due in the lecture the following Tuesday and will be returned and discussed on Thursday in the exercise course.

Exercise sheets:
  • Assignment 1     [pdf] [ps]
  • Assignment 2     [pdf] [ps]
  • Assignment 3     [pdf] [ps]
  • Assignment 4     [pdf] [ps]
  • Assignment 5     [pdf] [ps]
  • Assignment 6     [pdf] [ps]
  • Assignment 7     [pdf] [ps]
  • Assignment 8     [pdf] [ps]
  • Assignment 9     [pdf] [ps]
  • Assignment 10     [pdf] [ps]
  • Assignment 11     [pdf] [ps]


Forum: A discussion forum on anything related to the course can be found here.

Lecture details:
Date Lecturer Topic Reference
Oct 17 FN Introduction, RLS and (1+1) EA, ONEMAX, TRAP Notes [pdf] [ps]
Oct 24 BD Fitness-based partitions, Chernoff, coupon collector Notes [pdf]
Oct 31 NH Quality of EA and specialized algorithms Notes [pdf]
Nov 7 FN Design of EAs (representation, crossover, mutation, selection) Notes [pdf]
Nov 14 FN (1+1) EA and the minimum spanning tree problem
Nov 21 NH Minimum spanning tree problem - lower bound
Nov 28 no lecture, but talk of Eckart Zitzler on Nov 27, 16:15, R. 024 more information
Dec 5 NH SEMO and FEMO Notes [pdf]
Dec 12 NH Differences of (1+1) EA and (1+1)*EA on a plateau
Dec 19 FN Eulerian cycles
[christmas holidays]
Jan 9 BD Fast Solution to the Eulerian Cycle Problem
Jan 16 NH Metropolis Algorithm and Simulated Annealing I
Jan 23 NH Metropolis Algorithm and Simulated Annealing II
Jan 30 FN Ant Colony Optimization
Feb 6 FN ACO and Minimum Spanning Trees
Feb 13 NH


Literature: We suggest the following literature:

  • T. Baeck, D. B. Fogel, Z. Michalewicz: Handbook of Evolutionary Computation, IOP Publishing and Oxford University Press, New York, 1997
  • A.E. Eiben, J.E. Smith: Introduction to Evolutionary Computing, Springer, 2003
  • R. Motwani, P. Raghavan: Randomized Algorithms, Cambridge University Press, 1995
  • Z. Michalewicz, D. B. Fogel: How to Solve It: Modern Heuristics, Springer, 2004
  • F. Neumann: Combinatorial optimization and the analysis of randomized search heuristics (Ph. D. thesis) (Download)
  • Ingo Wegeners page (with script), SS 2002 [in german only]