combinatorial operads rewrite systems and formal grammars
play

Combinatorial operads, rewrite systems, and formal grammars Samuele - PowerPoint PPT Presentation

Combinatorial operads, rewrite systems, and formal grammars Samuele Giraudo LIGM, Universit Paris-Est Marne-la-Valle Computational Logic and Applications July 12, 2019 1 / 40 Outline Operads Enumeration Generation 2 / 40 Outline


  1. Some suboperads of Path For any m � 0 , an m -Dyck path is a path starting and ending with 0 and m made of steps and . 0 — Example — is a 2 -Dyck path of size 10 . — Proposition — For any m � 0 , the set Dyck ( m ) of all m -Dyck paths is a suboperad of Path . A Motzkin path is a path starting and ending with 0 and made of steps , , and . — Example — is a Motzkin path of size 16 . 10 / 40

  2. Some suboperads of Path For any m � 0 , an m -Dyck path is a path starting and ending with 0 and m made of steps and . 0 — Example — is a 2 -Dyck path of size 10 . — Proposition — For any m � 0 , the set Dyck ( m ) of all m -Dyck paths is a suboperad of Path . A Motzkin path is a path starting and ending with 0 and made of steps , , and . — Example — is a Motzkin path of size 16 . — Proposition — The set Motz of all Motzkin paths is a suboperad of Path . 10 / 40

  3. Algebras over operads Let O be an operad. An algebra over O is a space V equipped, for all x ∈ O ( n ) , with linear maps x : V ⊗ · · · ⊗ V → V � �� � n 11 / 40

  4. Algebras over operads Let O be an operad. An algebra over O is a space V equipped, for all x ∈ O ( n ) , with linear maps x : V ⊗ · · · ⊗ V → V � �� � n such that 1 is the identity map on V 11 / 40

  5. Algebras over operads Let O be an operad. An algebra over O is a space V equipped, for all x ∈ O ( n ) , with linear maps x : V ⊗ · · · ⊗ V → V � �� � n such that 1 is the identity map on V and the compatibility relation x x ◦ i y = . . . . . . v | x | + | y |− 1 v 1 y . . . v | x | + | y |− 1 v 1 . . . v i + | y |− 1 v i holds for any x, y ∈ O , i ∈ [ | x | ] , and v 1 , . . . , v | x | + | y |− 1 ∈ V . 11 / 40

  6. Algebras over operads — Example — Let As be the associative operad defined by As ( n ) := { ⋆ n } for all n � 1 and ⋆ n ◦ i ⋆ m := ⋆ n + m − 1 . 12 / 40

  7. Algebras over operads — Example — Let As be the associative operad defined by As ( n ) := { ⋆ n } for all n � 1 and ⋆ n ◦ i ⋆ m := ⋆ n + m − 1 . This operad is minimally generated by ⋆ 2 . 12 / 40

  8. Algebras over operads — Example — Let As be the associative operad defined by As ( n ) := { ⋆ n } for all n � 1 and ⋆ n ◦ i ⋆ m := ⋆ n + m − 1 . This operad is minimally generated by ⋆ 2 . Any algebra over As is a space V endowed with linear operations ⋆ n of arity n � 1 12 / 40

  9. Algebras over operads — Example — Let As be the associative operad defined by As ( n ) := { ⋆ n } for all n � 1 and ⋆ n ◦ i ⋆ m := ⋆ n + m − 1 . This operad is minimally generated by ⋆ 2 . Any algebra over As is a space V endowed with linear operations ⋆ n of arity n � 1 where ⋆ 2 satisfies, for all v 1 , v 2 , v 3 ∈ V , ( ⋆ 2 ◦ 1 ⋆ 2 ) ( v 1 , v 2 , v 3 ) � ( ⋆ 2 ◦ 2 ⋆ 2 ) ( v 1 , v 2 , v 3 ) 12 / 40

  10. Algebras over operads — Example — Let As be the associative operad defined by As ( n ) := { ⋆ n } for all n � 1 and ⋆ n ◦ i ⋆ m := ⋆ n + m − 1 . This operad is minimally generated by ⋆ 2 . Any algebra over As is a space V endowed with linear operations ⋆ n of arity n � 1 where ⋆ 2 satisfies, for all v 1 , v 2 , v 3 ∈ V , ( ⋆ 2 ◦ 1 ⋆ 2 ) ( v 1 , v 2 , v 3 ) = ⋆ 2 ( ⋆ 2 ( v 1 , v 2 ) , v 3 ) � ( ⋆ 2 ◦ 2 ⋆ 2 ) ( v 1 , v 2 , v 3 ) 12 / 40

  11. Algebras over operads — Example — Let As be the associative operad defined by As ( n ) := { ⋆ n } for all n � 1 and ⋆ n ◦ i ⋆ m := ⋆ n + m − 1 . This operad is minimally generated by ⋆ 2 . Any algebra over As is a space V endowed with linear operations ⋆ n of arity n � 1 where ⋆ 2 satisfies, for all v 1 , v 2 , v 3 ∈ V , ( ⋆ 2 ◦ 1 ⋆ 2 ) ( v 1 , v 2 , v 3 ) = ⋆ 2 ( ⋆ 2 ( v 1 , v 2 ) , v 3 ) � ( ⋆ 2 ◦ 2 ⋆ 2 ) ( v 1 , v 2 , v 3 ) = ⋆ 2 ( v 1 , ⋆ 2 ( v 2 , v 3 )) . 12 / 40

  12. Algebras over operads — Example — Let As be the associative operad defined by As ( n ) := { ⋆ n } for all n � 1 and ⋆ n ◦ i ⋆ m := ⋆ n + m − 1 . This operad is minimally generated by ⋆ 2 . Any algebra over As is a space V endowed with linear operations ⋆ n of arity n � 1 where ⋆ 2 satisfies, for all v 1 , v 2 , v 3 ∈ V , ( ⋆ 2 ◦ 1 ⋆ 2 ) ( v 1 , v 2 , v 3 ) = ⋆ 2 ( ⋆ 2 ( v 1 , v 2 ) , v 3 ) � � ( ⋆ 2 ◦ 2 ⋆ 2 ) ( v 1 , v 2 , v 3 ) = ⋆ 2 ( v 1 , ⋆ 2 ( v 2 , v 3 )) . 12 / 40

  13. Algebras over operads — Example — Let As be the associative operad defined by As ( n ) := { ⋆ n } for all n � 1 and ⋆ n ◦ i ⋆ m := ⋆ n + m − 1 . This operad is minimally generated by ⋆ 2 . Any algebra over As is a space V endowed with linear operations ⋆ n of arity n � 1 where ⋆ 2 satisfies, for all v 1 , v 2 , v 3 ∈ V , ( ⋆ 2 ◦ 1 ⋆ 2 ) ( v 1 , v 2 , v 3 ) = ⋆ 2 ( ⋆ 2 ( v 1 , v 2 ) , v 3 ) � � ( ⋆ 2 ◦ 2 ⋆ 2 ) ( v 1 , v 2 , v 3 ) = ⋆ 2 ( v 1 , ⋆ 2 ( v 2 , v 3 )) . Using infix notation for the binary operation ⋆ 2 , we obtain the relation ( v 1 ⋆ 2 v 2 ) ⋆ 2 v 3 = v 1 ⋆ 2 ( v 2 ⋆ 2 v 3 ) , so that algebras over As are associative algebras. 12 / 40

  14. Algebras over operads — Example — Let As be the associative operad defined by As ( n ) := { ⋆ n } for all n � 1 and ⋆ n ◦ i ⋆ m := ⋆ n + m − 1 . This operad is minimally generated by ⋆ 2 . Any algebra over As is a space V endowed with linear operations ⋆ n of arity n � 1 where ⋆ 2 satisfies, for all v 1 , v 2 , v 3 ∈ V , ( ⋆ 2 ◦ 1 ⋆ 2 ) ( v 1 , v 2 , v 3 ) = ⋆ 2 ( ⋆ 2 ( v 1 , v 2 ) , v 3 ) � � ( ⋆ 2 ◦ 2 ⋆ 2 ) ( v 1 , v 2 , v 3 ) = ⋆ 2 ( v 1 , ⋆ 2 ( v 2 , v 3 )) . Using infix notation for the binary operation ⋆ 2 , we obtain the relation ( v 1 ⋆ 2 v 2 ) ⋆ 2 v 3 = v 1 ⋆ 2 ( v 2 ⋆ 2 v 3 ) , so that algebras over As are associative algebras. In the same way, there are operads for ◮ Lie alg.; ◮ duplicial alg. [ Loday , 2008] ; ◮ pre-Lie alg. [ Chapoton, Livernet , 2001] ; ◮ diassociative alg. [ Loday , 2001] ; ◮ dendriform alg. [ Loday , 2001] ; ◮ brace alg. 12 / 40

  15. Scope of operads As main benefits, operads ◮ offer a formalism to compute over operations; ◮ allow us to work virtually with all the structures of a type; ◮ lead to discover the underlying combinatorics of types of algebras. 13 / 40

  16. Scope of operads As main benefits, operads ◮ offer a formalism to compute over operations; ◮ allow us to work virtually with all the structures of a type; ◮ lead to discover the underlying combinatorics of types of algebras. Endowing a set of combinatorial objects with an operad structure helps to ◮ highlight elementary building block for the objects; ◮ build combinatorial structures (graded graphs, posets, latices, etc. ); ◮ enumerative prospects and discovery of statistics. 13 / 40

  17. Outline Enumeration 14 / 40

  18. Syntax trees An alphabet is a graded set G := � n � 1 G ( n ) . 15 / 40

  19. Syntax trees An alphabet is a graded set G := � n � 1 G ( n ) . Let S ( G ) be the set of G -syntax trees, defined recursively by ◮ ∈ S ( G ) ; ◮ if a ∈ G and t 1 , . . . , t | a | ∈ S ( G ) , then a ( t 1 , . . . , t | a | ) ∈ S ( G ) . — Example — Let G := G (2) ⊔ G (3) such that G (2) = { a , b } and G (3) = { c } . denotes the G -tree c c b c ( , c ( a ( , ) , , b ( a ( , ) , c ( , , ) ) ) , b ( , b ( , ) ) ) a b b having degree 8 and arity 12 . a c 15 / 40

  20. Syntax trees An alphabet is a graded set G := � n � 1 G ( n ) . Let S ( G ) be the set of G -syntax trees, defined recursively by ◮ ∈ S ( G ) ; ◮ if a ∈ G and t 1 , . . . , t | a | ∈ S ( G ) , then a ( t 1 , . . . , t | a | ) ∈ S ( G ) . Let t ∈ S ( G ) . Some definitions: ◮ is the leaf; ◮ the degree deg( t ) of t is its number of internal nodes; ◮ the arity | t | of t is its number of leaves. — Example — Let G := G (2) ⊔ G (3) such that G (2) = { a , b } and G (3) = { c } . denotes the G -tree c c b c ( , c ( a ( , ) , , b ( a ( , ) , c ( , , ) ) ) , b ( , b ( , ) ) ) a b b having degree 8 and arity 12 . a c 15 / 40

  21. Compositions of syntax trees Let t , s ∈ S ( G ) . For each i ∈ [ | t | ] , the partial composition t ◦ i s is the tree obtained by grafing the root of s onto the i th leaf of t . — Example — c c c b b ◦ 5 = c b a a c b b a b a c 16 / 40

  22. Compositions of syntax trees Let t , s ∈ S ( G ) . For each i ∈ [ | t | ] , the partial composition t ◦ i s is the tree obtained by grafing the root of s onto the i th leaf of t . — Example — c c c b b ◦ 5 = c b a a c b b a b a c � � Let t , s 1 , ..., s | t | be G -trees. The full composition t ◦ s 1 , . . . , s | t | is obtained by grafing simultaneously the roots of each s i onto the i th leaf of t . — Example —   b a b ◦  =  b , ,  a a c  a a a c b 16 / 40

  23. Free operads Let G be an alphabet. 17 / 40

  24. Free operads Let G be an alphabet. The free operad on G is the operad on the set S ( G ) wherein ◮ elements of arity n are the G -trees of arity n ; 17 / 40

  25. Free operads Let G be an alphabet. The free operad on G is the operad on the set S ( G ) wherein ◮ elements of arity n are the G -trees of arity n ; ◮ the partial composition map ◦ i is the one of the G -trees; 17 / 40

  26. Free operads Let G be an alphabet. The free operad on G is the operad on the set S ( G ) wherein ◮ elements of arity n are the G -trees of arity n ; ◮ the partial composition map ◦ i is the one of the G -trees; ◮ the unit is . 17 / 40

  27. Free operads Let G be an alphabet. The free operad on G is the operad on the set S ( G ) wherein ◮ elements of arity n are the G -trees of arity n ; ◮ the partial composition map ◦ i is the one of the G -trees; ◮ the unit is . Let c : G → S ( G ) be the natural injection (made implicit in the sequel). 17 / 40

  28. Free operads Let G be an alphabet. The free operad on G is the operad on the set S ( G ) wherein ◮ elements of arity n are the G -trees of arity n ; ◮ the partial composition map ◦ i is the one of the G -trees; ◮ the unit is . Let c : G → S ( G ) be the natural injection (made implicit in the sequel). Free operads satisfy the following universality property. f For any alphabet G , any operad O , O G and any map f : G → O preserving the arities, there exists a unique op- c φ erad morphism φ : S ( G ) → O such that f = φ ◦ c . S ( G ) 17 / 40

  29. Factors and prefixes Let t , s ∈ S ( G ) . 18 / 40

  30. Factors and prefixes Let t , s ∈ S ( G ) . If t decomposes as � � �� t = r ◦ i s ◦ r 1 , . . . , r [ s | for some trees r , r 1 , ..., r | s | , and i ∈ [ | r | ] , then s is a factor of t . This property is denoted by s � f t . — Example — b c c b � f b a b a c b 18 / 40

  31. Factors and prefixes Let t , s ∈ S ( G ) . If t decomposes as � � �� t = r ◦ i s ◦ r 1 , . . . , r [ s | for some trees r , r 1 , ..., r | s | , and i ∈ [ | r | ] , then s is a factor of t . This property is denoted by s � f t . If in the previous decomposition r = , then s is a prefix of t . This property is denoted by s � p t . — Example — b b c c b b c b � f � p c b a b a c a a c b b b b 18 / 40

  32. Patern avoidance and enumeration ✚ A G -tree t avoids a G -tree s if s ✚ � f t . 19 / 40

  33. Patern avoidance and enumeration ✚ A G -tree t avoids a G -tree s if s ✚ � f t . For any P ⊆ S ( G ) , let A( P ) = { t ∈ S ( G ) : for all s ∈ P , s ✚ � f t } . ✚ 19 / 40

  34. Patern avoidance and enumeration ✚ A G -tree t avoids a G -tree s if s ✚ � f t . For any P ⊆ S ( G ) , let A( P ) = { t ∈ S ( G ) : for all s ∈ P , s ✚ � f t } . ✚ — Qestion — Enumerate A( P ) w.r.t. the arities of the trees. 19 / 40

  35. Patern avoidance and enumeration ✚ A G -tree t avoids a G -tree s if s ✚ � f t . For any P ⊆ S ( G ) , let A( P ) = { t ∈ S ( G ) : for all s ∈ P , s ✚ � f t } . ✚ — Example — � � ◮ A a a b b is enumerated by 1 , 2 , 4 , 8 , 16 , 32 , 64 , 128 , . . . . a b a b — Qestion — Enumerate A( P ) w.r.t. the arities of the trees. 19 / 40

  36. Patern avoidance and enumeration ✚ A G -tree t avoids a G -tree s if s ✚ � f t . For any P ⊆ S ( G ) , let A( P ) = { t ∈ S ( G ) : for all s ∈ P , s ✚ � f t } . ✚ — Example — � � ◮ A a a b b is enumerated by 1 , 2 , 4 , 8 , 16 , 32 , 64 , 128 , . . . . a b a b � � ◮ A a c c a is enumerated by 1 , 1 , 2 , 4 , 9 , 21 , 51 , 127 , . . . ( A001006 ) . c a c a — Qestion — Enumerate A( P ) w.r.t. the arities of the trees. 19 / 40

  37. Patern avoidance and enumeration ✚ A G -tree t avoids a G -tree s if s ✚ � f t . For any P ⊆ S ( G ) , let A( P ) = { t ∈ S ( G ) : for all s ∈ P , s ✚ � f t } . ✚ — Example — � � ◮ A a a b b is enumerated by 1 , 2 , 4 , 8 , 16 , 32 , 64 , 128 , . . . . a b a b � � ◮ A a c c a is enumerated by 1 , 1 , 2 , 4 , 9 , 21 , 51 , 127 , . . . ( A001006 ) . c a c a   b ◮ A  is enumerated by 1 , 2 , 5 , 13 , 35 , 96 , 267 , 750 ,. . . ( A005773 ) . a b b  b a a a b — Qestion — Enumerate A( P ) w.r.t. the arities of the trees. 19 / 40

  38. Formal power series of trees For any P , Q ⊆ S ( G ) , let � F ( P , Q ) := t . t ∈ S ( G ) t ∈ A( P ) � p t ✚ ∀ s ∈Q , s ✚ This is the formal sum of all the G -trees avoiding as factors all paterns of P and avoiding as prefixes all paterns of Q . 20 / 40

  39. Formal power series of trees For any P , Q ⊆ S ( G ) , let � F ( P , Q ) := t . t ∈ S ( G ) t ∈ A( P ) � p t ✚ ∀ s ∈Q , s ✚ This is the formal sum of all the G -trees avoiding as factors all paterns of P and avoiding as prefixes all paterns of Q . Since ◮ F ( P , ∅ ) is the formal sum of all the trees of A( P ) ; 20 / 40

  40. Formal power series of trees For any P , Q ⊆ S ( G ) , let � F ( P , Q ) := t . t ∈ S ( G ) t ∈ A( P ) � p t ✚ ∀ s ∈Q , s ✚ This is the formal sum of all the G -trees avoiding as factors all paterns of P and avoiding as prefixes all paterns of Q . Since ◮ F ( P , ∅ ) is the formal sum of all the trees of A( P ) ; ◮ the linear map t �→ z | t | sends F ( P , ∅ ) to the generating series of A( P ) ; 20 / 40

  41. Formal power series of trees For any P , Q ⊆ S ( G ) , let � F ( P , Q ) := t . t ∈ S ( G ) t ∈ A( P ) � p t ✚ ∀ s ∈Q , s ✚ This is the formal sum of all the G -trees avoiding as factors all paterns of P and avoiding as prefixes all paterns of Q . Since ◮ F ( P , ∅ ) is the formal sum of all the trees of A( P ) ; ◮ the linear map t �→ z | t | sends F ( P , ∅ ) to the generating series of A( P ) ; the series F ( P , Q ) contains all the enumerative data about the trees avoiding P . 20 / 40

  42. System of equations When G , P , and Q satisfy some conditions, F ( P , Q ) expresses as an inclusion-exclusion formula involving simpler terms F ( P , S i ) . — Theorem — The series F ( P , Q ) satisfies � � ( − 1) 1+ ℓ a ¯ F ( P , Q ) = + ◦ [ F ( P , S 1 ) , . . . , F ( P , S k )] . k � 1 ℓ � 1 a ∈ G ( k ) � R (1) ,..., R ( ℓ ) � ⊆ M (( P∪Q ) a ) ( S 1 ,..., S k )= R (1) ∔ ··· ∔ R ( ℓ ) 21 / 40

  43. System of equations When G , P , and Q satisfy some conditions, F ( P , Q ) expresses as an inclusion-exclusion formula involving simpler terms F ( P , S i ) . — Theorem — The series F ( P , Q ) satisfies � � ( − 1) 1+ ℓ a ¯ F ( P , Q ) = + ◦ [ F ( P , S 1 ) , . . . , F ( P , S k )] . k � 1 ℓ � 1 a ∈ G ( k ) � R (1) ,..., R ( ℓ ) � ⊆ M (( P∪Q ) a ) ( S 1 ,..., S k )= R (1) ∔ ··· ∔ R ( ℓ ) This leads to a system of equations for the generating series of A( P ) . Indeed, the generating series of A( P ) is the series F ( P , ∅ ) where � � ( − 1) 1+ ℓ � F ( P , Q ) = z + F ( P , S i ) . k � 1 ℓ � 1 i ∈ [ k ] a ∈ G ( k ) � R (1) ,..., R ( ℓ ) � ⊆ M (( P∪Q ) a ) ( S 1 ,..., S k )= R (1) ∔ ··· ∔ R ( ℓ ) 21 / 40

  44. System of equations — Example — � � For P := a , we obtain the system of formal power series of trees a b F ( P , ∅ ) = + a ¯ ◦ [ F ( P , { a } ) , F ( P , ∅ )] + a ¯ ◦ [ F ( P , ∅ ) , F ( P , { b } )] − a ¯ ◦ [ F ( P , { a } ) , F ( P , { b } )] + b ¯ ◦ [ F ( P , ∅ ) , F ( P , ∅ )] , F ( P , { a } ) = + b ¯ ◦ [ F ( P , ∅ ) , F ( P , ∅ )] , F ( P , { b } ) = + a ¯ ◦ [ F ( P , { a } ) , F ( P , ∅ )] + a ¯ ◦ [ F ( P , ∅ ) , F ( P , { b } )] − a ¯ ◦ [ F ( P , { a } ) , F ( P , { b } )] . 22 / 40

  45. System of equations — Example — � � For P := a , we obtain the system of formal power series of trees a b F ( P , ∅ ) = + a ¯ ◦ [ F ( P , { a } ) , F ( P , ∅ )] + a ¯ ◦ [ F ( P , ∅ ) , F ( P , { b } )] − a ¯ ◦ [ F ( P , { a } ) , F ( P , { b } )] + b ¯ ◦ [ F ( P , ∅ ) , F ( P , ∅ )] , F ( P , { a } ) = + b ¯ ◦ [ F ( P , ∅ ) , F ( P , ∅ )] , F ( P , { b } ) = + a ¯ ◦ [ F ( P , { a } ) , F ( P , ∅ )] + a ¯ ◦ [ F ( P , ∅ ) , F ( P , { b } )] − a ¯ ◦ [ F ( P , { a } ) , F ( P , { b } )] . This leads to the system of generating series F ( P , ∅ ) = z + F ( P , { a } ) F ( P , ∅ ) + F ( P , ∅ ) F ( P , { b } ) − F ( P , { a } ) F ( P , { b } ) + F ( P , ∅ ) F ( P , ∅ ) , F ( P , { a } ) = z + F ( P , ∅ ) F ( P , ∅ ) , F ( P , { b } ) = z + F ( P , { a } ) F ( P , ∅ ) + F ( P , ∅ ) F ( P , { b } ) − F ( P , { a } ) F ( P , { b } ) . 22 / 40

  46. System of equations — Example — � � For P := a , we obtain the system of formal power series of trees a b F ( P , ∅ ) = + a ¯ ◦ [ F ( P , { a } ) , F ( P , ∅ )] + a ¯ ◦ [ F ( P , ∅ ) , F ( P , { b } )] − a ¯ ◦ [ F ( P , { a } ) , F ( P , { b } )] + b ¯ ◦ [ F ( P , ∅ ) , F ( P , ∅ )] , F ( P , { a } ) = + b ¯ ◦ [ F ( P , ∅ ) , F ( P , ∅ )] , F ( P , { b } ) = + a ¯ ◦ [ F ( P , { a } ) , F ( P , ∅ )] + a ¯ ◦ [ F ( P , ∅ ) , F ( P , { b } )] − a ¯ ◦ [ F ( P , { a } ) , F ( P , { b } )] . This leads to the system of generating series F ( P , ∅ ) = z + F ( P , { a } ) F ( P , ∅ ) + F ( P , ∅ ) F ( P , { b } ) − F ( P , { a } ) F ( P , { b } ) + F ( P , ∅ ) F ( P , ∅ ) , F ( P , { a } ) = z + F ( P , ∅ ) F ( P , ∅ ) , F ( P , { b } ) = z + F ( P , { a } ) F ( P , ∅ ) + F ( P , ∅ ) F ( P , { b } ) − F ( P , { a } ) F ( P , { b } ) . As a consequence, F ( P , ∅ ) satisfies z − F ( P , ∅ ) + (2 + z ) F ( P , ∅ ) 2 − F ( P , ∅ ) 3 + F ( P , ∅ ) 4 = 0 . 22 / 40

  47. Operads and presentations Let O be an operad. A congruence of O is an equivalence relation ≡ on O preserving the arities and such that x ≡ x ′ and y ≡ y ′ imply x ◦ i y ≡ x ′ ◦ i y ′ for all i ∈ [ | x | ] . 23 / 40

  48. Operads and presentations Let O be an operad. A congruence of O is an equivalence relation ≡ on O preserving the arities and such that x ≡ x ′ and y ≡ y ′ imply x ◦ i y ≡ x ′ ◦ i y ′ for all i ∈ [ | x | ] . A presentation of O is a pair ( G , ≡ ) such that G is an alphabet and ≡ is a congruence of O satisfying O ≃ S ( G ) / ≡ . 23 / 40

  49. Operads and presentations Let O be an operad. A congruence of O is an equivalence relation ≡ on O preserving the arities and such that x ≡ x ′ and y ≡ y ′ imply x ◦ i y ≡ x ′ ◦ i y ′ for all i ∈ [ | x | ] . A presentation of O is a pair ( G , ≡ ) such that G is an alphabet and ≡ is a congruence of O satisfying O ≃ S ( G ) / ≡ . — Example — The operad Motz admits the presentation ( G , ≡ ) where � � G := , 23 / 40

  50. Operads and presentations Let O be an operad. A congruence of O is an equivalence relation ≡ on O preserving the arities and such that x ≡ x ′ and y ≡ y ′ imply x ◦ i y ≡ x ′ ◦ i y ′ for all i ∈ [ | x | ] . A presentation of O is a pair ( G , ≡ ) such that G is an alphabet and ≡ is a congruence of O satisfying O ≃ S ( G ) / ≡ . — Example — The operad Motz admits the presentation ( G , ≡ ) where � � G := , and ≡ is the smallest operad congruence satisfying ◦ 1 ≡ ◦ 2 , ◦ 1 ≡ ◦ 2 , ◦ 1 ≡ ◦ 3 , ◦ 1 ≡ ◦ 3 . 23 / 40

  51. Operads and paterns Let O be an operad admiting a presentation ( G , ≡ ) . 24 / 40

  52. Operads and paterns Let O be an operad admiting a presentation ( G , ≡ ) . A basis of O is a subset B of S ( G ) such that for any [ t ] ≡ ∈ S ( G ) / ≡ , there exists a unique s ∈ [ t ] ≡ ∩ B . 24 / 40

  53. Operads and paterns Let O be an operad admiting a presentation ( G , ≡ ) . A basis of O is a subset B of S ( G ) such that for any [ t ] ≡ ∈ S ( G ) / ≡ , there exists a unique s ∈ [ t ] ≡ ∩ B . In most cases, B can be described as set of G -trees avoiding a subset P B of S ( G ) . 24 / 40

  54. Operads and paterns Let O be an operad admiting a presentation ( G , ≡ ) . A basis of O is a subset B of S ( G ) such that for any [ t ] ≡ ∈ S ( G ) / ≡ , there exists a unique s ∈ [ t ] ≡ ∩ B . In most cases, B can be described as set of G -trees avoiding a subset P B of S ( G ) . — Example — The set B , described as the set of G -trees avoiding � � P B := ◦ 1 ◦ 1 ◦ 1 ◦ 1 , , , , is a basis of Motz . 24 / 40

  55. Operads and paterns Let O be an operad admiting a presentation ( G , ≡ ) . A basis of O is a subset B of S ( G ) such that for any [ t ] ≡ ∈ S ( G ) / ≡ , there exists a unique s ∈ [ t ] ≡ ∩ B . In most cases, B can be described as set of G -trees avoiding a subset P B of S ( G ) . — Example — The set B , described as the set of G -trees avoiding � � P B := ◦ 1 ◦ 1 ◦ 1 ◦ 1 , , , , is a basis of Motz . Rewrite systems on G -trees are good tools to compute bases (we find terminating and confluent orientations ⇒ of ≡ ). 24 / 40

  56. Operads and enumeration Let X be a family of combinatorial objects we want enumerate. The approach using operads consists in 1. endowing X with the structure of an operad O X ; 25 / 40

  57. Operads and enumeration Let X be a family of combinatorial objects we want enumerate. The approach using operads consists in 1. endowing X with the structure of an operad O X ; 2. exhibiting a presentation ( G , ≡ ) of O X and a basis B ; 25 / 40

  58. Operads and enumeration Let X be a family of combinatorial objects we want enumerate. The approach using operads consists in 1. endowing X with the structure of an operad O X ; 2. exhibiting a presentation ( G , ≡ ) of O X and a basis B ; 3. computing the series F ( P B , ∅ ) where P B is a set of G -trees satisfying A ( P B ) = B . 25 / 40

  59. Operads and enumeration Let X be a family of combinatorial objects we want enumerate. The approach using operads consists in 1. endowing X with the structure of an operad O X ; 2. exhibiting a presentation ( G , ≡ ) of O X and a basis B ; 3. computing the series F ( P B , ∅ ) where P B is a set of G -trees satisfying A ( P B ) = B . — Example — To enumerate Motzkin paths (w.r.t. their sizes), we consider their operad structure Motz . 25 / 40

  60. Operads and enumeration Let X be a family of combinatorial objects we want enumerate. The approach using operads consists in 1. endowing X with the structure of an operad O X ; 2. exhibiting a presentation ( G , ≡ ) of O X and a basis B ; 3. computing the series F ( P B , ∅ ) where P B is a set of G -trees satisfying A ( P B ) = B . — Example — To enumerate Motzkin paths (w.r.t. their sizes), we consider their operad structure Motz . � � a c c Let a := , c := , and P := a . c a c a 25 / 40

  61. Operads and enumeration Let X be a family of combinatorial objects we want enumerate. The approach using operads consists in 1. endowing X with the structure of an operad O X ; 2. exhibiting a presentation ( G , ≡ ) of O X and a basis B ; 3. computing the series F ( P B , ∅ ) where P B is a set of G -trees satisfying A ( P B ) = B . — Example — To enumerate Motzkin paths (w.r.t. their sizes), we consider their operad structure Motz . � � a c c Let a := , c := , and P := a . c a c a We have F ( P , ∅ ) = + a ¯ ◦ [ F ( P , { a , c } ) , F ( P , ∅ )] + c ¯ ◦ [ F ( P , { a , c } ) , F ( P , ∅ ) , F ( P , ∅ )] , F ( P , { a , c } ) = , so that, the generating series of Motzkin paths satisfies F ( P , ∅ ) = z + z F ( P , ∅ ) + z F ( P , ∅ ) 2 . 25 / 40

  62. Outline Generation 26 / 40

  63. Context-free grammars Let A = V ⊔ T be a set where V is a set of variables and T is a set of terminal symbols. 27 / 40

  64. Context-free grammars Let A = V ⊔ T be a set where V is a set of variables and T is a set of terminal symbols. A rule is a pair ( x, v ) ∈ V × A ∗ . A set R of rules specifies a rewrite rule → on A ∗ by seting u x w → u v w for any u, w ∈ A ∗ provided that ( x, v ) ∈ R . 27 / 40

  65. Context-free grammars Let A = V ⊔ T be a set where V is a set of variables and T is a set of terminal symbols. A rule is a pair ( x, v ) ∈ V × A ∗ . A set R of rules specifies a rewrite rule → on A ∗ by seting u x w → u v w for any u, w ∈ A ∗ provided that ( x, v ) ∈ R . — Example — Let V := { x, y } , T := { a , b , c } , and R := { ( x, b ) , ( x, x a y ) , ( y, ac ) } . We have b xx → b x a yx → bba yx → bbaac x. 27 / 40

  66. Regular tree grammars Let V be a set of variables and T be an alphabet of terminal symbols. 28 / 40

  67. Regular tree grammars Let V be a set of variables and T be an alphabet of terminal symbols. A ( V , T ) -tree is a planar rooted tree where internal nodes are labeled on T and leaves are labeled on V . 28 / 40

  68. Regular tree grammars Let V be a set of variables and T be an alphabet of terminal symbols. A ( V , T ) -tree is a planar rooted tree where internal nodes are labeled on T and leaves are labeled on V . A rule is a pair ( x, t ) where x ∈ V and t is a ( V , T ) -tree. A set R of rules specifies a rewrite rule → on the set of all ( V , T ) -trees by seting → s s x t for any ( V , T ) -tree s having a leaf labeled by x , provided that ( x, t ) ∈ R . 28 / 40

  69. Regular tree grammars Let V be a set of variables and T be an alphabet of terminal symbols. A ( V , T ) -tree is a planar rooted tree where internal nodes are labeled on T and leaves are labeled on V . A rule is a pair ( x, t ) where x ∈ V and t is a ( V , T ) -tree. A set R of rules specifies a rewrite rule → on the set of all ( V , T ) -trees by seting → s s x t for any ( V , T ) -tree s having a leaf labeled by x , provided that ( x, t ) ∈ R . — Example — �� � �� � b Let V := { x, y } , T := { a , b } where | a | := 1 , | b | := 2 , and R := x, , y, . a x b y x y b We have b a a b → → . x x a a a b x x y x b y x 28 / 40

  70. General generation Objectives: ◮ Introduce generating systems for any kind of combinatorial objects; ◮ Retrieve the generation of words and of trees as special cases; ◮ Develop a toolbox for the enumeration of combinatorial objects. 29 / 40

  71. General generation Objectives: ◮ Introduce generating systems for any kind of combinatorial objects; ◮ Retrieve the generation of words and of trees as special cases; ◮ Develop a toolbox for the enumeration of combinatorial objects. — Key idea — Use colored operads, where ◮ colors play the role of variables and terminal symbols; ◮ Formal series on colored operad and their operations support enumeration. 29 / 40

  72. Colored operads Colored operads are algebraic structures formalizing the notion of partial operations and their composition. 30 / 40

  73. Colored operads Colored operads are algebraic structures formalizing the notion of partial operations and their composition. A colored operad is a quadruplet ( C , C , ◦ i , 1 c ) where 1. C is a finite set of colors; 30 / 40

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend


More recommend