Publication Details
Evolution of Generic Square Calculations in Cellular Automata
cellular automaton, transition function, evolutionary algorithm, square calculation
The paper deals with the design of uniform multi-state one-dimensional cellular automata using an evolutionary algorithm and their application to solve the problem of generic square calculations. The key idea is based on the representation of the transition functions for the automata, which utilises the concept of conditionally matching rules. This technique allows us to design complex cellular automata for which the conventional representations have failed. A study is proposed with various settings of the experimental system, which concerns the way of evaluating the candidate solutions, the number of cell states and the number of conditional rules of the transition functions. It is shown that various generic solutions for the square calculation can be obtained in one-dimensional cellular automata using local interactions of cells only. The results presented demonstrates an ability of the evolution to discover innovative solutions both from the view of complexity of the cellular autom aton and the number of steps needed to calculate the results in comparison with the known solution.
@INPROCEEDINGS{FITPUB11150, author = "Michal Bidlo", title = "Evolution of Generic Square Calculations in Cellular Automata", pages = "94--102", booktitle = "Proceedings of the 8th International Joint Conference on Computational Intelligence - Volume 3: ECTA", year = 2016, location = "Porto, PT", publisher = "SciTePress - Science and Technology Publications", ISBN = "978-989-758-201-1", doi = "10.5220/0006064800940102", language = "english", url = "https://www.fit.vut.cz/research/publication/11150" }