Loading...

Design of Local Rule for Cellular Automata Using Evolutionary Algorithms

Mousavi, Samane Sadat | 2010

603 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 40117 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Beigy, Hamid
  7. Abstract:
  8. Cellular automata is a model for physical systems that has homogenous and simple components. Simple components, which are called cells, have local interactions creating complicated global emergent of behaviour. In the field of cellular automata, there are two basic problems: forward and inverse problems. Characterizations of cellular automata rule are studied in forward problem, but in inverse problem, there exists a description of cellular automata and we should find a rule or a set of rules that satisfy the given description. This problem belongs to the class of NP problems and hence heuristic algorithms such as evolutionary algorithms have been used for solving it. Since rules space and configurations space have exponential growth, concise fitness evaluation in polynomial time is impossible. In this thesis, we propose two methods. In the first method, we use clustering algorithm for selecting more appropriate cellular automata configurations on which rules are evaluated. In the second method, we propose a method for conducting mutation. Experimental results show that the first method outperforms random selection of configurations and the second method increases convergence percentage of genetic algorithm.

  9. Keywords:
  10. Evolutionary Algorithm ; Genetic Algorithm ; Clustering ; Data Reduction ; Cellular Automata ; Evolutionary Cellular Automata

 Digital Object List

  • محتواي پايان نامه
  •   view

 Bookmark

No TOC