Applied Evolutionary Algorithms

Ac.Year:ukončen 2005/2006
Language of Instruction:Czech
Private info:http://www.fit.vutbr.cz/study/courses/EVA/private/
Completion:examination (written)
Type of
Guarantor:Schwarz Josef, doc. Ing., CSc. (DCSY)
Lecturer:Schwarz Josef, doc. Ing., CSc. (DCSY)
Faculty:Faculty of Information Technology BUT
Department:Department of Computer Systems FIT BUT
Artificial Intelligence (UIN), DITS
Learning objectives:
  Survey about actual optimization techniques and evolutionary algorithms for solution of complex, NP complete problems. To make familiar students with software tools for fast prototyping of evolutionary algorithms and learn how to solve typical complex tasks in engineering practice.
  Theoretical and practical foundation of evolutionary computation. Evolutionary algorithms using genetic algorithms, evolution strategies, evolution programming, genetic programming and classifiers as probabilistic search algorithms. Techniques for fast prototyping of genetic algorithms. Advanced estimation distribution algorithms (EDA). Synergy of evolutionary computation and fuzzy logic. Evolutionary algorithms in engineering application: artificial intelligence, knowledge base systems, VLSI design and multiprocessor scheduling.
Knowledge and skills required for the course:
  Basic knowledge of algorithm theory and their complexity. Basic terms from graph theory, artificial intelligence and probability theory. 
Learning outcomes and competencies:
  Ability of problem formulation for the solution on the base of evolutionary computation. Knowledge of methodology for fast prototyping of evolutionary optimizer utilizing GA library and current design tools.
Syllabus of lectures:
  • Evolutionary algorithms, basic classification. Optimization tasks. 
  • Genetic algorithms (GA), schema theorem.
  • Advanced genetic algorithms, diploids, messy-chromozomes.
  • Combinatorial tasks. Evolution strategies.
  • Evolution programming. Genetic programming.
  • Simulated annealing. Hill climbing algorithms. Tabu search.
  • Evolutionary algorithms with probabilistic models (EDA algorithms).
  • Variants of EDA algorithms - UMDA, BMDA, BOA.
  • Multimodal and multiobjective tasks.
  • Dynamical optimization tasks. Immune systems.
  • Hybrid genetic algorithms. Techniques for fast prototyping.
  • Synergy of genetic algorithms, fuzzy logic and neural networks. Classifiers.
  • Typical problems in engineering practice.
Syllabus of laboratory exercises:
  • EA - Introduction
  • Presentation of the GADesign design tool.
  • Optimizer design with the GADesign tool.
  • Structure and utilizing of GALIB library.
  • Genetic programming.
  • Presentation of the DEBOA design tool.
Syllabus - others, projects and individual work of students:
  • Program implementation for the solution of a given optimization  problem by means of the evolutionary algorithms.
Fundamental literature:
  • Eiben, A. E., Smith, E.: Introduction to Evolutionary Computing (Natural Computing Series). Springer Verlag, November, 2003, pp. 299, ISBN 3540401849.
  • Dasgupta, D., Michalewicz, Z.: Evolutionary algorithms in engineering applications. Springer Verlag, Berlin, 1997, ISBN 3-540-62021-4.
  • Back, J: Evolutionary algorithms, theory and practice, New York, 1996.
  • Kvasnička, V., Pospíchal, J.,Tiňo, P.: Evolutionary algorithms. Publisher STU Bratislava, 2000, pp. 215, ISBN 80-227-1377-5.
Study literature:
  • Kvasnička V., Pospíchal J., Tiňo P.: Evolutionary algorithms. Publisher STU Bratislava, 2000, pp. 215, ISBN 80-227-1377-5.
  • Kvasnička V., a kol.: Introduction into theory of neural networks, Iris 1997, ISBN 80-88778-30-1.
Controlled instruction:
  Project is monitored
Progress assessment:
  • Mid-term written exam - 20 points.
  • Individual project - 30 points.
  • Final written examination - 50 points.
  • Passing boundary for ECTS assessment - 50 points.
Exam prerequisites:
  Requirements for class accreditation are not defined.

Your IPv4 address:
Switch to https

DNSSEC [dnssec]