design and architectures for embedded systems
play

Design and Architectures for Embedded Systems Prof. Dr. J. Henkel - PowerPoint PPT Presentation

GA 1 Design and Architectures for Embedded Systems Prof. Dr. J. Henkel Prof. Dr. J. Henkel CES - - Chair for Embedded Systems Chair for Embedded Systems CES University of Karlsruhe, Germany University of Karlsruhe, Germany Add- -on


  1. GA 1 Design and Architectures for Embedded Systems Prof. Dr. J. Henkel Prof. Dr. J. Henkel CES - - Chair for Embedded Systems Chair for Embedded Systems CES University of Karlsruhe, Germany University of Karlsruhe, Germany Add- -on Slides: Genetic/Evolutionary Algorithm on Slides: Genetic/Evolutionary Algorithm Add http://ces.univ-karlsruhe.de J. Henkel, Univ. of Karlsruhe, WS0809

  2. GA 2 Genetic & Evolutionary Algorithms � Genetic Algorithms belong to the area of evolutionary computing � Genetic Algorithms belong to the area of evolutionary computing which which itself is part of AI itself is part of AI � Short history � Short history � 1960s: I. � 1960s: I. Rechenberg Rechenberg (“Evolution Strategies”) (“Evolution Strategies”) � 1975: Genetic Algorithms by John Holland “Adaption in natural an � 1975: Genetic Algorithms by John Holland “Adaption in natural and artificial d artificial systems” systems” � 1992: J. � 1992: J. Koza Koza, Genetic Programming , Genetic Programming � Background � Background � Rooted in the mechanism of evolution and natural genetics � Rooted in the mechanism of evolution and natural genetics � Draws inspirations from the natural search and selection process � Draws inspirations from the natural search and selection process - -> > “survival of the fittest” “survival of the fittest” � Based on sequences of the following mechanisms: � Based on sequences of the following mechanisms: � Selection � Selection � Crossover � Crossover � Mutation � Mutation http://ces.univ-karlsruhe.de J. Henkel, Univ. of Karlsruhe, WS0809

  3. GA 3 Selection A 1 B 1 A 2 B 1 B 2 . B 2 . . . . . . . . A N B N B N C 1 C 1 C 2 D 1 C 2 D 1 . D 2 . D 2 . . . . . . . . . . C N C N D N D N http://ces.univ-karlsruhe.de J. Henkel, Univ. of Karlsruhe, WS0809

  4. 4 GA http://ces.univ-karlsruhe.de C N C 1 C 2 . . . Crossover J. Henkel, Univ. of Karlsruhe, WS0809 B N A N B 1 B 2 A 1 A 2 . . . . . .

  5. GA 5 Mutation A 1 A 1 A 2 B 1 A 2 B 1 . B 2 . B 2 . . . . . . . . X . . A N A N B N B N C 1 C 1 C 2 D 1 C 2 D 1 . D 2 . D 2 . . . Y . . . . . . . C N C N D N D N http://ces.univ-karlsruhe.de J. Henkel, Univ. of Karlsruhe, WS0809

  6. GA 6 Simple generic GA Begin Initialize population; Evaluate population; While ( termination criterion is not satisfied ) Select solutions for next population; Perform crossover; Perform mutation; Evaluate population; End http://ces.univ-karlsruhe.de J. Henkel, Univ. of Karlsruhe, WS0809

  7. GA 7 Summary GA, EA � Resembles evolution in natural genetics � Resembles evolution in natural genetics � Difference GA < � Difference GA <- -> EA > EA � GA: use � GA: use crossover crossover as primary search strategy as primary search strategy � EA: use � EA: use mutation mutation as primary search strategy as primary search strategy � There is no mathematical foundation for how well GA/EA � There is no mathematical foundation for how well GA/EA optimize optimize � Try yourself … for example, the “Traveling Sales Man � Try yourself … for example, the “Traveling Sales Man Problem” … Problem” … http://ces.univ-karlsruhe.de J. Henkel, Univ. of Karlsruhe, WS0809

Recommend


More recommend