reasoning about connectivity without paths
play

Reasoning about connectivity without paths Alberto Casagrande and - PowerPoint PPT Presentation

Reasoning about connectivity without paths Alberto Casagrande and Eugenio G. Omodeo Dip. Matematica e Geoscienze DMI Eugenio G. Omodeo Reasoning about Connectivity without Paths 1/24 Reasoning about connectivity without paths 1 Alberto


  1. Reasoning about connectivity without paths Alberto Casagrande and Eugenio G. Omodeo Dip. Matematica e Geoscienze — DMI Eugenio G. Omodeo Reasoning about Connectivity without Paths 1/24

  2. Reasoning about connectivity without paths 1 Alberto Casagrande and Eugenio G. Omodeo Dip. Matematica e Geoscienze — DMI 1 Work partially funded by: INdAM/GNCS 2013, FRA-UniTS 2012 PUMA Eugenio G. Omodeo Reasoning about Connectivity without Paths 1/24

  3. Every connected graph has non-cut vertices Example Eugenio G. Omodeo Reasoning about Connectivity without Paths 2/24

  4. Every connected graph has non-cut vertices Example Eugenio G. Omodeo Reasoning about Connectivity without Paths 2/24

  5. Every connected graph has non-cut vertices Example Eugenio G. Omodeo Reasoning about Connectivity without Paths 2/24

  6. Every connected graph has non-cut vertices Example Eugenio G. Omodeo Reasoning about Connectivity without Paths 2/24

  7. Every connected graph has non-cut vertices Example · Eugenio G. Omodeo Reasoning about Connectivity without Paths 2/24

  8. Every connected graph has non-cut vertices Example · Eugenio G. Omodeo Reasoning about Connectivity without Paths 2/24

  9. Every connected graph has non-cut vertices Example More generally: ‘Every connected (finite) hyper graph has at least one vertex whose removal does not disrupt connectivity’ Eugenio G. Omodeo Reasoning about Connectivity without Paths 2/24

  10. (Ideal) itinerary of this talk i Connectivity and non-cut vertices ii Applications ( one in particular. . . ) iii The proof assistant Ref iv Our proof-verification experiment http://www2.units.it/eomodeo/NonCutVertices.html http://aetnanova.units.it/scenarios/NonCutVertices/ Eugenio G. Omodeo Reasoning about Connectivity without Paths 3/24

  11. Connectivity Connectivity plays a crucial role in many fields. Es. The number of connected components of a graph is a topological invariant; corresponds to the multiplicity of the eigenvalue 0 in its Laplacian; is related to the number of its claw-free subgraphs [CPR07]. ∴ Large scale proof-verification efforts [Wie07, SCO11] must formally investigate this notion. Eugenio G. Omodeo Reasoning about Connectivity without Paths 4/24

  12. How we see a hypergraph Example f a b g c h e d Eugenio G. Omodeo Reasoning about Connectivity without Paths 5/24

  13. How we see a hypergraph Example f a b g c h e d The edges of G belong to G . Eugenio G. Omodeo Reasoning about Connectivity without Paths 5/24

  14. How we see a hypergraph Example f a b g c h e d   vertices    of G belong to { v : e ∈ G , v ∈ e } . The or  nodes Eugenio G. Omodeo Reasoning about Connectivity without Paths 5/24

  15. How we see a hypergraph Example f a b g c h e d In a graph, the edges have cardinality 2 . Eugenio G. Omodeo Reasoning about Connectivity without Paths 5/24

  16. Our formal definition of connectivity . . . Def . Conn ( G ) { p ⊆ G | nodes ( p ) ∩ nodes ( G \ p ) = ∅ } ⊆ { ∅ , G } & ↔ Def HGraph ( G ) Eugenio G. Omodeo Reasoning about Connectivity without Paths 6/24

  17. Our formal definition of connectivity . Def . nodes ( G ) = Def � G . Def . Conn ( G ) { p ⊆ G | nodes ( p ) ∩ nodes ( G \ p ) = ∅ } = { ∅ , G } & ↔ Def HGraph ( G ) Eugenio G. Omodeo Reasoning about Connectivity without Paths 6/24

  18. Our formal definition of connectivity . Def . nodes ( G ) = Def � G � ∀ e ∈ G | CardAtLeast2 ( e ) � & Def . HGraph ( G ) ↔ Def � � Finite nodes ( G ) Def . Conn ( G ) { p ⊆ G | nodes ( p ) ∩ nodes ( G \ p ) = ∅ } = { ∅ , G } & ↔ Def HGraph ( G ) Eugenio G. Omodeo Reasoning about Connectivity without Paths 6/24

  19. Our formal definition of connectivity Def . CardAtLeast2 ( E ) ↔ Def E �⊆ { arb ( E ) } Def . nodes ( G ) = Def � G � ∀ e ∈ G | CardAtLeast2 ( e ) � & Def . HGraph ( G ) ↔ Def � � Finite nodes ( G ) Def . Conn ( G ) { p ⊆ G | nodes ( p ) ∩ nodes ( G \ p ) = ∅ } = { ∅ , G } & ↔ Def HGraph ( G ) Eugenio G. Omodeo Reasoning about Connectivity without Paths 6/24

  20. Does this hypergraph have a spanning tree ? c a e b f d Eugenio G. Omodeo Reasoning about Connectivity without Paths 7/24

  21. Existence of non-cut vertices . . . Thm . Conn ( G ) & G �⊆ { arb ( G ) } → � ∃ v ∈ nodes ( G ) | NonCut ( G , v ) � Eugenio G. Omodeo Reasoning about Connectivity without Paths 8/24

  22. Existence of non-cut vertices . . � � & lost ( G , V ) = ∅ Def . NonCut ( G , V ) Conn rmv ( G , V ) ↔ Def Thm . Conn ( G ) & G �⊆ { arb ( G ) } → � ∃ v ∈ nodes ( G ) | NonCut ( G , v ) � Eugenio G. Omodeo Reasoning about Connectivity without Paths 8/24

  23. Existence of non-cut vertices Def . rmv ( G , V ) = Def { e \ { V } : e ∈ G | CardAtLeast2 ( e \ { V } ) } . � � & lost ( G , V ) = ∅ Def . NonCut ( G , V ) Conn rmv ( G , V ) ↔ Def Thm . Conn ( G ) & G �⊆ { arb ( G ) } → � ∃ v ∈ nodes ( G ) | NonCut ( G , v ) � Eugenio G. Omodeo Reasoning about Connectivity without Paths 8/24

  24. Existence of non-cut vertices Def . rmv ( G , V ) = Def { e \ { V } : e ∈ G | CardAtLeast2 ( e \ { V } ) } � � � � Def . lost ( G , V ) = Def nodes ( G ) \ nodes rmv ( G , V ) ∪ { V } � � & lost ( G , V ) = ∅ Def . NonCut ( G , V ) Conn rmv ( G , V ) ↔ Def Thm . Conn ( G ) & G �⊆ { arb ( G ) } → � ∃ v ∈ nodes ( G ) | NonCut ( G , v ) � Eugenio G. Omodeo Reasoning about Connectivity without Paths 8/24

  25. Ur-application of non-cut vertices: Walking How can we: walk along an infinite acyclic path? visit all vertices of a finite acyclic path? Eugenio G. Omodeo Reasoning about Connectivity without Paths 9/24

  26. Ur-application of non-cut vertices: Walking Example 4 · · · 0 1 2 3 Eugenio G. Omodeo Reasoning about Connectivity without Paths 9/24

  27. Ur-application of non-cut vertices: Walking Example 4 · · · 0 1 2 3 Eugenio G. Omodeo Reasoning about Connectivity without Paths 9/24

  28. Ur-application of non-cut vertices: Walking Example 4 · · · 1 2 3 Eugenio G. Omodeo Reasoning about Connectivity without Paths 9/24

  29. Ur-application of non-cut vertices: Walking Example 4 · · · 2 3 Eugenio G. Omodeo Reasoning about Connectivity without Paths 9/24

  30. Ur-application of non-cut vertices: Walking Example 4 · · · 3 Eugenio G. Omodeo Reasoning about Connectivity without Paths 9/24

  31. Ur-application of non-cut vertices: Walking Example 4 · · · Eugenio G. Omodeo Reasoning about Connectivity without Paths 9/24

  32. Ur-application of non-cut vertices: Walking Example 4 · · · This amounts to repeatedly picking and removing a non-cut vertex (the only one , in this case) from a graph ( infinite in this case) Eugenio G. Omodeo Reasoning about Connectivity without Paths 9/24

  33. Ur-application of non-cut vertices: Walking Example 4 · · · This amounts to repeatedly picking and removing a non-cut vertex (the only one , in this case) from a graph ( infinite in this case) Why such a silly example ? Eugenio G. Omodeo Reasoning about Connectivity without Paths 9/24

  34. Ur-application of non-cut vertices: Walking Example 4 · · · 0 1 2 3 N ∪ { N } N Where does the difference between N and N ∪ { N , N ∪ { N }} lie ? Eugenio G. Omodeo Reasoning about Connectivity without Paths 9/24

  35. A more revealing application Example We can get a spanning tree for a connected nonnull graph by: Eugenio G. Omodeo Reasoning about Connectivity without Paths 10/24

  36. A more revealing application Example We can get a spanning tree for a connected nonnull graph by: 1 Picking & removing a non-cut vertex from a connected graph Eugenio G. Omodeo Reasoning about Connectivity without Paths 10/24

  37. A more revealing application Example We can get a spanning tree for a connected nonnull graph by: 1 Picking & removing a non-cut vertex from a connected graph recursively getting a spanning tree for the resulting graph 2 Eugenio G. Omodeo Reasoning about Connectivity without Paths 10/24

  38. A more revealing application Example We can get a spanning tree for a connected nonnull graph by: 1 Picking & removing a non-cut vertex from a connected graph recursively getting a spanning tree for the resulting graph 2 3 restoring the removed vertex, along with one of the edges incident to it Eugenio G. Omodeo Reasoning about Connectivity without Paths 10/24

  39. A more revealing application Example We can get a spanning tree for a connected nonnull graph by: 1 Picking & removing a non-cut vertex from a connected graph recursively getting a spanning tree for the resulting graph 2 3 restoring the removed vertex, along with one of the edges incident to it 0 In the base case , the spanning tree consists of the (sole) edge Eugenio G. Omodeo Reasoning about Connectivity without Paths 10/24

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