Course Computational Intelligence


Responsible: Prof. Dr. Rainer Bartz

Course

Meets requirements of following modules(MID)

Course Organization

Version
created 2013-06-20
VID 2
valid from WS 2012/13
valid to
Course identifiers
Long name Computational Intelligence
CID F07_CI
CEID (exam identifier)

Contact hours per week (SWS)
Lecture 2
Exercise (unsplit) 1
Exercise (split)
Lab 1
Project
Seminar
Tutorial(voluntary)
Total contact hours
Lecture 30
Exercise (unsplit) 15
Exercise (split)
Lab 15
Project
Seminar
Tutorial (voluntary)
Max. capacity
Exercise (unsplit) 30
Exercise (split)
Lab 10
Project
Seminar

Total effort (hours): 150

Instruction language

  • German, 40%
  • English, 60%

Study Level

  • graduate

Prerequisites

  • vector functions, gradient

Textbooks, Recommended Reading

  • Domschke W., Drexl A.; Einführung in Operations Research; Springer
  • Zell, A.: Simulation Neuronaler Netze; Oldenbourg
  • Nauck, D. et al.: Neuronale Netze und Fuzzy-Systeme; Vieweg
  • Eiben, A.E., Smith, J.E.: Introduction to Evolutionary Computing; Springer
  • Gerdes, I. et al.: Evolutionäre Algorithmen; Vieweg
  • Grosse et al.: Taschenbuch der praktischen Regelungstechnik, Fachbuchverlag Leipzig

Instructors

  • Prof. Dr. Rainer Bartz

Supporting Scientific Staff

  • tba

Transcipt Entry

Computational Intelligence

Assessment

Type
wE written exam

Total effort [hours]
wE 10

Frequency: 2/year


Course components

Lecture/Exercise

Objectives

Contents
  • optimization strategies
    • classification of problems
    • gradient algorithms
    • simplex algorithm
    • multiobjective optimization and Pareto approach
  • artificial neural networks
    • artificial neurons
    • neural network structures
    • training algorithms
  • fuzzy logic
    • fuzzification
    • inference
    • defuzzification
  • evolutionary algorithms
    • genome representations
    • selection mechanisms
    • recombination operators
    • mutation operators

Acquired Skills
  • students acquire fundamental knowledge on theory and applications of computational intelligence
  • they know about typical classes of optimization tasks and how to map a specific problem to those classes
  • they know the simplex algorithm and can transform problems into the standard form to find the solutions
  • they can classify artificial neural networks and determine their applicability for specific tasks
  • they can vary the parameters of neural networks and rate their impact on the results
  • they can classify training algorithms and understand the backpropagation algorithm
  • they know about the fuzzy logic approach, can apply it for specific problems and justify the resulting system behavior
  • they know how evolutionary algorithms work and can distinguish the variants
  • they can transform a problem specification into a representation appropriate for an evolutionary algorithm
  • they can rate selection strategies and define suitable algorithms

Operational Competences
  • students can solve linear problems with the use of the simplex algorithm
  • they can apply artificial neural networks to solve problems of modeling and classification
  • they can define fuzzy logic systems to solve imprecise and vague tasks
  • they can solve difficult problems heuristically using evolutionary algorithms

Additional Component Assessment

Type
fPS supervised/assisted problem solving

Contribution to course grade
fPS not rated

Frequency: 1/year

Lab

Objectives

Contents
  • application of artificial neural networks to a classification task
  • variation and multiobjective optimization of neural network parameters
  • fuzzy-based closed loop control of a system with two inputs

Acquired Skills
  • students are familiar with tools supporting computational intelligence
  • they can vary system parameters, perform test series, and evaluate, present and discuss the results
  • they are able to understand, present, analyze and discuss scientific publications

Operational Competences
  • students are able to solve problems in small teams
  • they can tackle optimization tasks in a structured and systematic way
  • they can rate the behavior of a system with regard to objectives and study and improve the behavior through parameter variations
  • they are able to cope with international scientific publications, understanding, presenting and discussing them in their context

Additional Component Assessment

Type
fSC 2 lab experiments
fLP supervised literature study and presentation

Contribution to course grade
fSC prerequisite for course exam
fLP prerequisite for course exam

Frequency: 1/year

Diese Seite läuft auf FoswikiDas Urheberrecht © liegt bei den mitwirkenden Autoren. Alle Inhalte dieser Kollaborations-Plattform sind Eigentum der Autoren.
Ideen, Anfragen oder Probleme bezüglich Foswiki? Feedback senden