wir bieten...
Dekobild im Seitenkopf ISMLL
Courses in winter term 2008/2009 / Project Artificial Intelligence:

Time: 14:15
Location: B026 Spl.
Begin: Mo. 27.10.08

Artificial Iintelligence is concerned with automation of intelligent behavior. The main techniques in the Artificial Intelligence can be divided into the following groups: "Searching" , "Planing" , "Optimization Techniques” , "Reasoning" and "Approximation Methods". Artificial intelligence is used in many applications, it is used for example in search engines, in expert systems, in the analysis and forecast of stock prices and in handwriting and speech recognition application. In the offered project the students will have the possibility to apply AI techniques in an specific given application.


Each group needs to give two presentations :

  • The first presentation is to show the progress of the work
    (First implementation and problems)
  • The second presentation is about the final results of the

Registration :

  • Registration is possible starting by now by email (Christine Preisach)
  • .
  • Topics will be assigned in the order of registration.
  • If you propose several topics, you will get the first assigned,
    which is not already taken.
  • Registration from groups, which already exist is desired.


Sudoku is a popular game for some years already. Sudoku can be interpreted as
a Constraint Satisfaction Problem (CSP) and solved with standard CSP techniques.

The goal of this project is the development of a software the solves Sudoku

The following subtasks have to be solved:
  1. Search for state of the art methods for solving Sudoku
  2. Implementation of a solver with different solution strategies
  3. Empirical comparison of the different strategies on the basis of
    example Sudokus, in particular on the basis of particularly difficult
    instances; additionally comparison of the developed program to also
    publicly available existing solver.
Qualitative reasoning
In qualitative reasoning in contrary to quantitative reasoning, not numeric procedures (e.g. Solving of linear equations), but symbolic computations are used in order to gain information about different> entities in space and time and of their relations to each other. Such a symbolic representation and computation can be advantageous regarding comprehensibility and computation complexity, because with it quantitative/metric data has not to be considered. Areas of application are automatic planning, geographical information systems and the semantic modelling of natural-language expressions.
The goal of this project is the development of a program for solving Allens Interval Algebra, which is a classical formalism of qualitative reasoning. The following subtasks could be solved:
  1. Design and implementation of a simple solver for Allens Interval Algebra
  2. Formalization of Allens Interval Algebra in first order logic and comparison of the efficiency of the combination formalization+FOL solver (e.g. SPASS) with the developed solver
  3. Implementation of a graphical user interface to be able to easily create and modify constraint networks in Allens Interval Algebra
Solve the Rubik's Cube with a Roboter
Construction of intelligent machines, robots is one of the oldest dreams of artifical intelligence. Intelligence is often understood as rational behaviour in various situations, applications, as strategy in "games" or as the ability of solving problems, puzzles.
The Rubik's Cube is one of the classic puzzles. Since Erno Rubik introduced his famous cube, various strategies, algorithms have been introduced to solve it. In this project student has to implement a solution for this puzzle using a Lego Mindstorms roboter.
The task consits of the followings:
  1. An overview of different solver algorithms (strategies) for the Rubik's Cube, and their qualitative comparision
  2. Implementation of some of the solver strategies as a prototype on a PC, and an empirical, quantitatieve comparision on their complexity (runtime needed to calculate the solution)
  3. Implementation of at least one of the solver strategies on a Lego Mindstorms robot. So that the robot works like on this page: http://tiltedtwister.com/index.html
(The Lego Mindstorms robot is provided for this task.)