The tour of a spanning tree Start from the root. Iterate the move : External half-edge = ⇒ Turn . Internal half-edge = ⇒ Cross + turn . Example : CRM, October 2006 ▽ Olivier Bernardi - CRM – p.15/31
The tour of a spanning tree Start from the root. Iterate the move : External half-edge = ⇒ Turn . Internal half-edge = ⇒ Cross + turn . Example : CRM, October 2006 ▽ Olivier Bernardi - CRM – p.15/31
The tour of a spanning tree Start from the root. Iterate the move : External half-edge = ⇒ Turn . Internal half-edge = ⇒ Cross + turn . Example : CRM, October 2006 ▽ Olivier Bernardi - CRM – p.15/31
The tour of a spanning tree Start from the root. Iterate the move : External half-edge = ⇒ Turn . Internal half-edge = ⇒ Cross + turn . Example : CRM, October 2006 ▽ Olivier Bernardi - CRM – p.15/31
The tour of a spanning tree Start from the root. Iterate the move : External half-edge = ⇒ Turn . Internal half-edge = ⇒ Cross + turn . Example : CRM, October 2006 ▽ Olivier Bernardi - CRM – p.15/31
The tour of a spanning tree Start from the root. Iterate the move : External half-edge = ⇒ Turn . Internal half-edge = ⇒ Cross + turn . Example : CRM, October 2006 ▽ Olivier Bernardi - CRM – p.15/31
The tour of a spanning tree Start from the root. Iterate the move : External half-edge = ⇒ Turn . Internal half-edge = ⇒ Cross + turn . Example : CRM, October 2006 ▽ Olivier Bernardi - CRM – p.15/31
The tour of a spanning tree Start from the root. Iterate the move : External half-edge = ⇒ Turn . Internal half-edge = ⇒ Cross + turn . Example : CRM, October 2006 Olivier Bernardi - CRM – p.15/31
Tour: graphical interpretation CRM, October 2006 Olivier Bernardi - CRM – p.16/31
Activities of spanning trees CRM, October 2006 ▽ Olivier Bernardi - CRM – p.17/31
Activities of spanning trees Consider the order of appearance of the edges around the tree. CRM, October 2006 ▽ Olivier Bernardi - CRM – p.17/31
Activities of spanning trees 5 6 3 4 2 1 Consider the order of appearance of the edges around the tree. CRM, October 2006 ▽ Olivier Bernardi - CRM – p.17/31
Activities of spanning trees 5 6 3 4 2 1 Consider the order of appearance of the edges around the tree. An internal/external edge is active if it is minimal in its fundamental cocycle/cycle. CRM, October 2006 ▽ Olivier Bernardi - CRM – p.17/31
Activities of spanning trees 5 6 3 4 2 1 Consider the order of appearance of the edges around the tree. We denote by: • i ( A ) the number of internal active edges, • e ( A ) the number of external active edges. CRM, October 2006 Olivier Bernardi - CRM – p.17/31
Tutte polynomial Theorem [B.] : The Tutte polynomial is equal to : � x i ( T ) y e ( T ) . T G ( x, y ) = T spanning tree CRM, October 2006 ▽ Olivier Bernardi - CRM – p.18/31
Tutte polynomial Theorem [B.] : The Tutte polynomial is equal to : � x i ( T ) y e ( T ) . T G ( x, y ) = T spanning tree Example : T K 3 ( x, y ) = x 2 + x + y. x 2 T K 3 ( x, y ) = + + x y CRM, October 2006 ▽ Olivier Bernardi - CRM – p.18/31
Tutte polynomial Theorem [B.] : The Tutte polynomial is equal to : � x i ( T ) y e ( T ) . T G ( x, y ) = T spanning tree Subgraph expansion [Brylawsky] : � ( x − 1) c ( S ) − 1 ( y − 1) | S | + c ( S ) −| V | . T G ( x, y ) = S ⊆ G CRM, October 2006 Olivier Bernardi - CRM – p.18/31
From spanning trees to subgraphs CRM, October 2006 ▽ Olivier Bernardi - CRM – p.19/31
From spanning trees to subgraphs CRM, October 2006 ▽ Olivier Bernardi - CRM – p.19/31
From spanning trees to subgraphs T + T T − CRM, October 2006 Olivier Bernardi - CRM – p.19/31
Partition of the subgraphs 2 E = � [ T − ; T + ] Proposition [B.] : T spanning tree CRM, October 2006 ▽ Olivier Bernardi - CRM – p.20/31
Partition of the subgraphs 2 E = � [ T − ; T + ] Proposition [B.] : T spanning tree Example : CRM, October 2006 ▽ Olivier Bernardi - CRM – p.20/31
Partition of the subgraphs 2 E = � [ T − ; T + ] Proposition [B.] : T spanning tree Example : CRM, October 2006 ▽ Olivier Bernardi - CRM – p.20/31
Partition of the subgraphs 2 E = � [ T − ; T + ] Proposition [B.] : T spanning tree Proposition [B.] : ( x − 1) c ( S ) − 1 ( y − 1) | S | + c ( S ) −| V | = x i ( T ) y e ( T ) � S ∈ [ T − ,T + ] CRM, October 2006 Olivier Bernardi - CRM – p.20/31
Bijection subgraphs ⇐ ⇒ orientations CRM, October 2006 Olivier Bernardi - CRM – p.21/31
Φ : Subgraphs �→ Orientations For spanning trees: CRM, October 2006 ▽ Olivier Bernardi - CRM – p.22/31
Φ : Subgraphs �→ Orientations For spanning trees: Internal edges are oriented from father to son. CRM, October 2006 ▽ Olivier Bernardi - CRM – p.22/31
Φ : Subgraphs �→ Orientations For spanning trees: Internal edges are oriented from father to son. External edges are oriented in such a way their heads appear before their tails around the tree. CRM, October 2006 ▽ Olivier Bernardi - CRM – p.22/31
Φ : Subgraphs �→ Orientations For spanning trees: Internal edges are oriented from father to son. External edges are oriented in such a way their heads appear before their tails around the tree. CRM, October 2006 Olivier Bernardi - CRM – p.22/31
Activity et acyclicity Remark : The fundamental cycle of an external active edge is oriented. CRM, October 2006 ▽ Olivier Bernardi - CRM – p.23/31
Activity et acyclicity Remark : The fundamental cycle of an external active edge is oriented. Hence, Φ( T ) acyclic = ⇒ T internal. CRM, October 2006 ▽ Olivier Bernardi - CRM – p.23/31
Activity et acyclicity Remark : The fundamental cycle of an external active edge is oriented. Hence, Φ( T ) acyclic ⇐ ⇒ T internal. CRM, October 2006 Olivier Bernardi - CRM – p.23/31
Activity et strong connectivity Remark : The fundamental cocycle of an internal active edge is oriented. CRM, October 2006 ▽ Olivier Bernardi - CRM – p.24/31
Activity et strong connectivity Remark : The fundamental cocycle of an internal active edge is oriented. Hence, Φ( T ) strongly connected = ⇒ T external. CRM, October 2006 ▽ Olivier Bernardi - CRM – p.24/31
Activity et strong connectivity Remark : The fundamental cocycle of an internal active edge is oriented. Hence, Φ( T ) strongly connected ⇐ ⇒ T external. CRM, October 2006 Olivier Bernardi - CRM – p.24/31
Φ : Subgraphs �→ Orientations For any subgraph S : Φ ? CRM, October 2006 ▽ Olivier Bernardi - CRM – p.25/31
Φ : Subgraphs �→ Orientations For any subgraph S : The subgraph S is in an interval [ T − , T + ] . Φ ? Φ CRM, October 2006 ▽ Olivier Bernardi - CRM – p.25/31
Φ : Subgraphs �→ Orientations For any subgraph S : The subgraph S is in an interval [ T − , T + ] . Reverse the fundamental cycle/cocycle of the edges in T △ S . Φ Φ CRM, October 2006 Olivier Bernardi - CRM – p.25/31
Φ : Subgraphs �→ Orientations Theorem [B.] : For any graph G , the mapping Φ establishes a bijection between subgraphs and orientations. CRM, October 2006 ▽ Olivier Bernardi - CRM – p.26/31
Φ : Subgraphs �→ Orientations Theorem [B.] : For any graph G , the mapping Φ establishes a bijection between subgraphs and orientations. Example : CRM, October 2006 ▽ Olivier Bernardi - CRM – p.26/31
Φ : Subgraphs �→ Orientations Theorem [B.] : For any graph G , the mapping Φ establishes a bijection between subgraphs and orientations. Example : CRM, October 2006 Olivier Bernardi - CRM – p.26/31
Specializations Subgraphs Orientations general root- connected strongly connected general connectedexternal general mal general forest mini internal acyclic CRM, October 2006 Olivier Bernardi - CRM – p.27/31
Specializations Connected: T G (1 , 2) CRM, October 2006 ▽ Olivier Bernardi - CRM – p.28/31
Specializations Connected: T G (1 , 2) :Root-connected . CRM, October 2006 ▽ Olivier Bernardi - CRM – p.28/31
Specializations External: T G (0 , 2) CRM, October 2006 ▽ Olivier Bernardi - CRM – p.28/31
Specializations External: T G (0 , 2) :Strongly-connected . CRM, October 2006 ▽ Olivier Bernardi - CRM – p.28/31
Specializations Forests: T G (2 , 1) CRM, October 2006 ▽ Olivier Bernardi - CRM – p.28/31
Specializations Forests: T G (2 , 1) :Minimal ( ⇐ ⇒ score vectors). CRM, October 2006 ▽ Olivier Bernardi - CRM – p.28/31
Specializations Internal: T G (2 , 0) CRM, October 2006 ▽ Olivier Bernardi - CRM – p.28/31
Specializations Internal: T G (2 , 0) :Acyclic . CRM, October 2006 Olivier Bernardi - CRM – p.28/31
Specializations Theorem [B.] : For any graph G , the mapping Φ induces a bijection between: CRM, October 2006 ▽ Olivier Bernardi - CRM – p.29/31
Specializations Theorem [B.] : For any graph G , the mapping Φ induces a bijection between: • Connected subgraphs and root-connected orientations. T G (1 , 2) [Gioan 06] CRM, October 2006 ▽ Olivier Bernardi - CRM – p.29/31
Specializations Theorem [B.] : For any graph G , the mapping Φ induces a bijection between: • Connected subgraphs and root-connected orientations. T G (1 , 2) [Gioan 06] • Forests and minimal orientations ( ⇐ ⇒ score vectors). T G (2 , 1) [Stanley 80] CRM, October 2006 ▽ Olivier Bernardi - CRM – p.29/31
Specializations Theorem [B.] : For any graph G , the mapping Φ induces a bijection between: • Connected subgraphs and root-connected orientations. T G (1 , 2) [Gioan 06] • Forests and minimal orientations ( ⇐ ⇒ score vectors). T G (2 , 1) [Stanley 80] • Trees and minimal root-connected orientations ( ⇐ ⇒ root-connected score vectors). T G (1 , 1) [Gioan 06] CRM, October 2006 Olivier Bernardi - CRM – p.29/31
Theorem [B.] : For any graph G , the mapping Φ induces a bijection between: • Internal subgraphs and acyclic orientations. T G (2 , 0) [Winder 66, Stanley 73, Gessel & Sagan 96] CRM, October 2006 ▽ Olivier Bernardi - CRM – p.30/31
Theorem [B.] : For any graph G , the mapping Φ induces a bijection between: • Internal subgraphs and acyclic orientations. T G (2 , 0) [Winder 66, Stanley 73, Gessel & Sagan 96] • External subgraphs and strongly-connected orientations. T G (0 , 2) [Las Vergnas 84, Gioan & Las Vergnas 06] CRM, October 2006 ▽ Olivier Bernardi - CRM – p.30/31
Theorem [B.] : For any graph G , the mapping Φ induces a bijection between: • Internal subgraphs and acyclic orientations. T G (2 , 0) [Winder 66, Stanley 73, Gessel & Sagan 96] • External subgraphs and strongly-connected orientations. T G (0 , 2) [Las Vergnas 84, Gioan & Las Vergnas 06] • Internal trees and acyclic root-connected orientations. T G (1 , 0) [Greene & Zaslavsky 83, Gessel & Sagan 96, Gebhard & Sagan 00] CRM, October 2006 ▽ Olivier Bernardi - CRM – p.30/31
Theorem [B.] : For any graph G , the mapping Φ induces a bijection between: • Internal subgraphs and acyclic orientations. T G (2 , 0) [Winder 66, Stanley 73, Gessel & Sagan 96] • External subgraphs and strongly-connected orientations. T G (0 , 2) [Las Vergnas 84, Gioan & Las Vergnas 06] • Internal trees and acyclic root-connected orientations. T G (1 , 0) [Greene & Zaslavsky 83, Gessel & Sagan 96, Gebhard & Sagan 00] • External trees and minimal strongly-connected orientations ( ⇐ ⇒ strongly-connected score vectors). T G (0 , 1) [Gioan 06] CRM, October 2006 Olivier Bernardi - CRM – p.30/31
Recommend
More recommend