max plus automata and tropical identities
play

Max-plus automata and Tropical identities Laure Daviaud University - PowerPoint PPT Presentation

Max-plus automata and Tropical identities Laure Daviaud University of Warwick Birmingham, 15-11-2017 Matrices vs machines... 2/15 Matrices vs machines... Matrices over ( N {} , max , +) 0 1


  1. Max-plus automata and Tropical identities Laure Daviaud University of Warwick Birmingham, 15-11-2017

  2. Matrices vs machines... 2/15

  3. Matrices vs machines... Matrices over ( N ∪ {−∞} , max , +)   0 −∞ −∞ −∞ 1 −∞   −∞ −∞ 0   0 0 −∞ −∞ −∞ 0   −∞ −∞ 0 2/15

  4. Matrices vs machines... Matrices over ( N ∪ {−∞} , max , +) Max-plus Automata   0 −∞ −∞ a , b : 0 a , b : 0 a : 1 −∞ 1 −∞   b : 0 b : 0 −∞ −∞ 0   0 0 −∞ −∞ −∞ 0   −∞ −∞ 0 2/15

  5. A very simple machine: Automata Finite alphabet A = { a , b } Set of words A ∗ : finite sequences of a and b 3/15

  6. A very simple machine: Automata Finite alphabet A = { a , b } Set of words A ∗ : finite sequences of a and b Check if a word has at least two b ’s. 3/15

  7. A very simple machine: Automata Finite alphabet A = { a , b } Set of words A ∗ : finite sequences of a and b Check if a word has at least two b ’s. a , b a , b a b b 3/15

  8. A very simple machine: Automata Finite alphabet A = { a , b } Set of words A ∗ : finite sequences of a and b Check if a word has at least two b ’s. a , b a , b a b b A word is accepted by the automaton if there is a path labelled by the word from an initial state to a final state. In this case [ [ A ] ]( w ) = 0, otherwise [ [ A ] ]( w ) = −∞ . 3/15

  9. A very simple machine: Automata Finite alphabet A = { a , b } Set of words A ∗ : finite sequences of a and b Check if a word has at least two b ’s. a , b a , b a b b A word is accepted by the automaton if there is a path labelled by the word from an initial state to a final state. In this case [ [ A ] ]( w ) = 0, otherwise [ [ A ] ]( w ) = −∞ . − → Quantitative extension: Weighted automata [Schützenberger, 61] 3/15

  10. Max-plus automata a , b : 0 a , b : 0 a : 1 b : 0 b : 0 4/15

  11. Max-plus automata a , b : 0 a , b : 0 a : 1 b : 0 b : 0 Syntax : Non deterministic finite automaton for which each transition is labelled by a non negative integer (weight). 4/15

  12. Max-plus automata a , b : 0 a , b : 0 a : 1 b : 0 b : 0 Syntax : Non deterministic finite automaton for which each transition is labelled by a non negative integer (weight). Semantic : Weight of a run: sum of the weights of the transitions. A + → N ∪ {−∞} w �→ Max of the weights of the accepting runs labelled by w ( −∞ if no such run) 4/15

  13. Max-plus automata a , b : 0 a , b : 0 a : 1 b : 0 b : 0 Syntax : Non deterministic finite automaton for which each transition is labelled by a non negative integer (weight). Semantic : Weight of a run: sum of the weights of the transitions. A + → N ∪ {−∞} w �→ Max of the weights of the accepting runs labelled by w ( −∞ if no such run) a n 0 ba n 1 b · · · ba n k +1 �→ max( n 1 , . . . , n k ) 4/15

  14. Matrix representation a , b : 0 a , b : 0 a : 1 b : 0 b : 0 5/15

  15. Matrix representation a , b : 0 a , b : 0 a : 1 b : 0 b : 0   0 −∞ −∞ µ ( a ) = −∞ 1 −∞   −∞ −∞ 0 5/15

  16. Matrix representation a , b : 0 a , b : 0 a : 1 b : 0 b : 0     0 −∞ −∞ 0 0 −∞ µ ( a ) = −∞ 1 −∞ µ ( b ) = −∞ −∞ 0     −∞ −∞ 0 −∞ −∞ 0 5/15

  17. Matrix representation a , b : 0 a , b : 0 a : 1 b : 0 b : 0     0 −∞ −∞ 0 0 −∞ µ ( a ) = −∞ 1 −∞ µ ( b ) = −∞ −∞ 0     −∞ −∞ 0 −∞ −∞ 0   −∞ � � I = 0 −∞ −∞ F = −∞   0 5/15

  18. Matrix representation a , b : 0 a , b : 0 a : 1 b : 0 b : 0     0 −∞ −∞ 0 0 −∞ µ ( a ) = −∞ 1 −∞ µ ( b ) = −∞ −∞ 0     −∞ −∞ 0 −∞ −∞ 0   −∞ � � I = 0 −∞ −∞ F = −∞   0 µ ( w ) i , j = max of the weights of the runs from i to j labelled by w [ [ A ] ]( w ) = I µ ( w ) F 5/15

  19. Matrix representation a , b : 0 a , b : 0 a : 1 b : 0 b : 0     0 −∞ −∞ 0 0 −∞ µ ( a ) = −∞ 1 −∞ µ ( b ) = −∞ −∞ 0     −∞ −∞ 0 −∞ −∞ 0   −∞ � � I = 0 −∞ −∞ F = −∞   0 µ ( w ) i , j = max of the weights of the runs from i to j labelled by w [ [ A ] ]( w ) = I µ ( w ) F Dimension = Number of states 5/15

  20. Questions ? Decidability and complexity � Equivalence [Krob] � Boundedness [Simon] � Determinisation [Kirsten, Klimann, Lombardy, Mairesse, Prieur] � Minimisation � ... 6/15

  21. A natural and fundamental question: A [ [ A ] ]( u ) u = ? A [ [ A ] ]( v ) v Which pairs of inputs can be distinguished by a given computational model? 7/15

  22. Distinguishing words Semiring ( N ∪ {−∞} , max , +) ] : A ∗ → N ∪ {−∞} [ [ A ] � � [ [ A ] ] : w �→ max ρ 1 + ρ 2 + · · · + ρ | w | ρ accepting path labelled by w C : class of the max-plus automata 8/15

  23. Distinguishing words Semiring ( N ∪ {−∞} , max , +) ] : A ∗ → N ∪ {−∞} [ [ A ] � � [ [ A ] ] : w �→ max ρ 1 + ρ 2 + · · · + ρ | w | ρ accepting path labelled by w C : class of the max-plus automata 1 For all u � = v , is there A ∈ C which distinguishes u and v ? → Yes 8/15

  24. Distinguishing words Semiring ( N ∪ {−∞} , max , +) ] : A ∗ → N ∪ {−∞} [ [ A ] � � [ [ A ] ] : w �→ max ρ 1 + ρ 2 + · · · + ρ | w | ρ accepting path labelled by w C : class of the max-plus automata 1 For all u � = v , is there A ∈ C which distinguishes u and v ? → Yes Is there A ∈ C which distinguishes all pairs u � = v ? 2 → No 8/15

  25. Distinguishing words Semiring ( N ∪ {−∞} , max , +) ] : A ∗ → N ∪ {−∞} [ [ A ] � � [ [ A ] ] : w �→ max ρ 1 + ρ 2 + · · · + ρ | w | ρ accepting path labelled by w C : class of the max-plus automata 1 For all u � = v , is there A ∈ C which distinguishes u and v ? → Yes Is there A ∈ C which distinguishes all pairs u � = v ? 2 → No 3 Minimal size to distinguish two given input words? → ?????? 8/15

  26. Given a positive integer n , are there u � = v such that for all max-plus automata A with at most n states: [ [ A ] ]( u ) = [ [ A ] ]( v ) ? 9/15

  27. Given a positive integer n , are there u � = v such that for all max-plus automata A with at most n states: [ [ A ] ]( u ) = [ [ A ] ]( v ) ? For matrices: Given a dimension n , does there exists a non trivial identity for the semigroup of square matrices of dimension n ? 9/15

  28. If n = 1 A = { a , b } a : α b : β 10/15

  29. If n = 1 A = { a , b } a : α b : β w �→ α | w | a + β | w | b 10/15

  30. If n = 1 A = { a , b } a : α b : β w �→ α | w | a + β | w | b Max-plus automata with one state can distinguish words with different contents (in particular different lengths), and only these ones. 10/15

  31. If n = 2 or n = 3 There exist pairs of distinct words with the same values for all automata with at most 3 states... 11/15

  32. If n = 2 or n = 3 There exist pairs of distinct words with the same values for all automata with at most 3 states... 2 states [Izhakian, Margolis] - words of length 20 11/15

  33. If n = 2 or n = 3 There exist pairs of distinct words with the same values for all automata with at most 3 states... 2 states [Izhakian, Margolis] - words of length 20 3 states [Shitov] - words of length 1795308 11/15

  34. Triangular automata 12/15

  35. Triangular automata Theorem [Izhakian] For all n , there exist a pair of distinct words u � = v such that for all triangular automata A with at most n states, [ [ A ] ]( u ) = [ [ A ] ]( v ) 12/15

  36. Triangular automata Theorem [Izhakian] For all n , there exist a pair of distinct words u � = v such that for all triangular automata A with at most n states, [ [ A ] ]( u ) = [ [ A ] ]( v ) For n = 2, exactly the identities for the bicyclic monoid [D., Johnson, Kambites] 12/15

  37. Let’s go back to automata with 2 states A = { a , b } a : α 1 a : α 3 a : α 2 b : β 2 a : α 4 b : β 4 b : β 1 b : β 3 13/15

  38. Let’s go back to automata with 2 states A = { a , b } a : α 1 a : α 3 a : α 2 b : β 2 a : α 4 b : β 4 b : β 1 b : β 3 Theorem [D., Johnson] - counter-example to a conjecture of Izhakian There are two pairs of distinct words of minimal length which cannot be distinguished by any max-plus automata with two states: a 2 b 3 a 3 babab 3 a 2 = a 2 b 3 ababa 3 b 3 a 2 and ab 3 a 4 baba 2 b 3 a = ab 3 a 2 baba 4 b 3 a 13/15

  39. Let’s go back to automata with 2 states A = { a , b } a : α 1 a : α 3 a : α 2 b : β 2 a : α 4 b : β 4 b : β 1 b : β 3 14/15

  40. Let’s go back to automata with 2 states A = { a , b } a : α 1 a : α 3 a : α 2 b : β 2 a : α 4 b : β 4 b : β 1 b : β 3 First attempt: Restrict the class of automata we have to consider 14/15

  41. Let’s go back to automata with 2 states A = { a , b } a : α 1 a : α 3 a : α 2 b : β 2 a : α 4 b : β 4 b : β 1 b : β 3 First attempt: Restrict the class of automata we have to consider � R − → Q − → Z − → N 14/15

Recommend


More recommend