an improved lp based approximation for steiner tree
play

An Improved LP-Based Approximation for Steiner Tree Fabrizio - PowerPoint PPT Presentation

Mar 15, 2023 •170 likes •1.21k views

An Improved LP-Based Approximation for Steiner Tree Fabrizio Grandoni Tor Vergata Rome grandoni@disp.uniroma2.it Joint work with J. Byrka, T. Rothvo, L. Sanit` a p. 1/29 The Steiner Tree Problem Def (Steiner tree) Given an undirected

  1. Algorithm IRR • For t = 1 , 2 , . . . ⋄ Compute a (1 + ε ) -apx solution x t for DCR ⋄ Sample a component C = C t with probability C / � p t C := x t D ∈C x t D ⋄ Contract C t and update DCR consequently ⋄ If there is only one terminal, output the sampled components Rem By adding a dummy component in the root, we can assume w.l.o.g. that M := � D ∈C x t D is fixed for all t – p. 12/29

  2. Bridge Lemma – p. 13/29

  3. Bridges Def Given a Steiner tree S and R ′ ⊆ R , the bridges br S,c ( R ′ ) of S w.r.t. R ′ (and edge costs c ) are the edges of S which do not belong to the minimum spanning tree of V ( S ) after the contraction of R ′ S 1 R ′ 8 10 1 2 9 2 1 – p. 14/29

  4. Bridges Def Given a Steiner tree S and R ′ ⊆ R , the bridges br S,c ( R ′ ) of S w.r.t. R ′ (and edge costs c ) are the edges of S which do not belong to the minimum spanning tree of V ( S ) after the contraction of R ′ 0 0 0 0 S 1 R ′ 8 10 1 2 9 2 1 – p. 14/29

  5. Bridges Def Given a Steiner tree S and R ′ ⊆ R , the bridges br S,c ( R ′ ) of S w.r.t. R ′ (and edge costs c ) are the edges of S which do not belong to the minimum spanning tree of V ( S ) after the contraction of R ′ 0 0 0 0 S R ′ 1 2 9 2 1 – p. 14/29

  6. Bridges Def Given a Steiner tree S and R ′ ⊆ R , the bridges br S,c ( R ′ ) of S w.r.t. R ′ (and edge costs c ) are the edges of S which do not belong to the minimum spanning tree of V ( S ) after the contraction of R ′ S 1 R ′ 8 10 br S,c ( R ′ ) 1 2 9 2 1 – p. 14/29

  7. Bridges Def Given a Steiner tree S and R ′ ⊆ R , the bridges br S,c ( R ′ ) of S w.r.t. R ′ (and edge costs c ) are the edges of S which do not belong to the minimum spanning tree of V ( S ) after the contraction of R ′ S 1 R ′ 8 10 br S,c ( R ′ ) 1 2 9 2 1 Rem The most expensive edge on a path between two gray nodes is a bridge – p. 14/29

  8. Bridges Def Given a Steiner tree S and R ′ ⊆ R , the bridges br S,c ( R ′ ) of S w.r.t. R ′ (and edge costs c ) are the edges of S which do not belong to the minimum spanning tree of V ( S ) after the contraction of R ′ S 1 R ′ 8 10 br S,c ( R ′ ) 1 2 9 2 1 Rem Let br S ( R ′ ) = br S,c ( R ′ ) , br S ( R ′ ) := c ( br S ( R ′ )) and br S ( C ) := br S ( R ∩ C ) . – p. 14/29

  9. Bridges Lem For any Steiner tree S on R , br S ( R ) ≥ 1 2 c ( S ) 5 1 2 4 1 3 – p. 15/29

  10. The Bridge Lemma Lem (Bridge Lemma) For any terminal spanning tree T and any feasible fractional solution x to DCR, � C ∈C x C · br T ( C ) ≥ c ( T ) – p. 16/29

  11. The Bridge Lemma Lem (Bridge Lemma) For any terminal spanning tree T and any feasible fractional solution x to DCR, � C ∈C x C · br T ( C ) ≥ c ( T ) • For every C ∈ C , with capacity x C , construct a directed terminal spanning tree Y C on R ∩ C , with capacity x C and edge weights w , as follows 1 8 10 1 2 9 2 1 – p. 16/29

  12. The Bridge Lemma Lem (Bridge Lemma) For any terminal spanning tree T and any feasible fractional solution x to DCR, � C ∈C x C · br T ( C ) ≥ c ( T ) • For every C ∈ C , with capacity x C , construct a directed terminal spanning tree Y C on R ∩ C , with capacity x C and edge weights w , as follows 1 8 10 1 2 9 2 1 – p. 16/29

  13. The Bridge Lemma Lem (Bridge Lemma) For any terminal spanning tree T and any feasible fractional solution x to DCR, � C ∈C x C · br T ( C ) ≥ c ( T ) • For every C ∈ C , with capacity x C , construct a directed terminal spanning tree Y C on R ∩ C , with capacity x C and edge weights w , as follows 1 2 9 2 1 – p. 16/29

  14. The Bridge Lemma Lem (Bridge Lemma) For any terminal spanning tree T and any feasible fractional solution x to DCR, � C ∈C x C · br T ( C ) ≥ c ( T ) • For every C ∈ C , with capacity x C , construct a directed terminal spanning tree Y C on R ∩ C , with capacity x C and edge weights w , as follows 8 1 2 9 2 1 – p. 16/29

  15. The Bridge Lemma Lem (Bridge Lemma) For any terminal spanning tree T and any feasible fractional solution x to DCR, � C ∈C x C · br T ( C ) ≥ c ( T ) • For every C ∈ C , with capacity x C , construct a directed terminal spanning tree Y C on R ∩ C , with capacity x C and edge weights w , as follows 8 8 1 2 9 2 1 – p. 16/29

  16. The Bridge Lemma Lem (Bridge Lemma) For any terminal spanning tree T and any feasible fractional solution x to DCR, � C ∈C x C · br T ( C ) ≥ c ( T ) • For every C ∈ C , with capacity x C , construct a directed terminal spanning tree Y C on R ∩ C , with capacity x C and edge weights w , as follows 8 10 1 2 9 2 1 – p. 16/29

  17. The Bridge Lemma Lem (Bridge Lemma) For any terminal spanning tree T and any feasible fractional solution x to DCR, � C ∈C x C · br T ( C ) ≥ c ( T ) • For every C ∈ C , with capacity x C , construct a directed terminal spanning tree Y C on R ∩ C , with capacity x C and edge weights w , as follows 10 8 10 1 2 9 2 1 – p. 16/29

  18. The Bridge Lemma Lem (Bridge Lemma) For any terminal spanning tree T and any feasible fractional solution x to DCR, � C ∈C x C · br T ( C ) ≥ c ( T ) • For every C ∈ C , with capacity x C , construct a directed terminal spanning tree Y C on R ∩ C , with capacity x C and edge weights w , as follows 10 8 1 1 2 9 2 1 – p. 16/29

  19. The Bridge Lemma Lem (Bridge Lemma) For any terminal spanning tree T and any feasible fractional solution x to DCR, � C ∈C x C · br T ( C ) ≥ c ( T ) • For every C ∈ C , with capacity x C , construct a directed terminal spanning tree Y C on R ∩ C , with capacity x C and edge weights w , as follows 10 8 1 1 2 9 2 1 – p. 16/29

  20. The Bridge Lemma Lem (Bridge Lemma) For any terminal spanning tree T and any feasible fractional solution x to DCR, � C ∈C x C · br T ( C ) ≥ c ( T ) • For every C ∈ C , with capacity x C , construct a directed terminal spanning tree Y C on R ∩ C , with capacity x C and edge weights w , as follows 10 8 1 1 2 9 2 1 – p. 16/29

  21. The Bridge Lemma Lem (Bridge Lemma) For any terminal spanning tree T and any feasible fractional solution x to DCR, � C ∈C x C · br T ( C ) ≥ c ( T ) • For every C ∈ C , with capacity x C , construct a directed terminal spanning tree Y C on R ∩ C , with capacity x C and edge weights w , as follows 10 8 1 1 2 9 2 1 Rem Y C supports the same flow to the root as C w.r.t. terminals – p. 16/29

  22. The Bridge Lemma Lem (Bridge Lemma) For any terminal spanning tree T and any feasible fractional solution x to DCR, � C ∈C x C · br T ( C ) ≥ c ( T ) 4 3 2 – p. 17/29

  23. The Bridge Lemma Lem (Bridge Lemma) For any terminal spanning tree T and any feasible fractional solution x to DCR, � C ∈C x C · br T ( C ) ≥ c ( T ) 4 3 2 – p. 17/29

  24. The Bridge Lemma Lem (Bridge Lemma) For any terminal spanning tree T and any feasible fractional solution x to DCR, � C ∈C x C · br T ( C ) ≥ c ( T ) 4 3 2 • Replace each component C with the corresponding Y C (cumulating capacities) – p. 17/29

  25. The Bridge Lemma Lem (Bridge Lemma) For any terminal spanning tree T and any feasible fractional solution x to DCR, � C ∈C x C · br T ( C ) ≥ c ( T ) 4 4 3 3 2 • Replace each component C with the corresponding Y C (cumulating capacities) – p. 17/29

  26. The Bridge Lemma Lem (Bridge Lemma) For any terminal spanning tree T and any feasible fractional solution x to DCR, � C ∈C x C · br T ( C ) ≥ c ( T ) 4 4 4 3 3 2 3 • Replace each component C with the corresponding Y C (cumulating capacities) – p. 17/29

  27. The Bridge Lemma Lem (Bridge Lemma) For any terminal spanning tree T and any feasible fractional solution x to DCR, � C ∈C x C · br T ( C ) ≥ c ( T ) 4 4 4 4 2 3 3 2 3 • Replace each component C with the corresponding Y C (cumulating capacities) – p. 17/29

  28. The Bridge Lemma Lem (Bridge Lemma) For any terminal spanning tree T and any feasible fractional solution x to DCR, � C ∈C x C · br T ( C ) ≥ c ( T ) 4 4 4 2 3 3 • We obtain a feasible fractional directed terminal spanning tree on a directed graph with V = R and edge costs w – p. 17/29

  29. The Bridge Lemma Lem (Bridge Lemma) For any terminal spanning tree T and any feasible fractional solution x to DCR, � C ∈C x C · br T ( C ) ≥ c ( T ) 4 4 4 4 2 3 2 3 3 • We obtain a feasible fractional directed terminal spanning tree on a directed graph with V = R and edge costs w ⇒ By Edmod’s thr there is a cheaper (w.r.t. w ) integral directed terminal spanning tree F – p. 17/29

  30. The Bridge Lemma Lem (Bridge Lemma) For any terminal spanning tree T and any feasible fractional solution x to DCR, � C ∈C x C · br T ( C ) ≥ c ( T ) 4 2 3 • We obtain a feasible fractional directed terminal spanning tree on a directed graph with V = R and edge costs w ⇒ By Edmod’s thr there is a cheaper (w.r.t. w ) integral directed terminal spanning tree F – p. 17/29

  31. The Bridge Lemma Lem (Bridge Lemma) For any terminal spanning tree T and any feasible fractional solution x to DCR, � C ∈C x C · br T ( C ) ≥ c ( T ) 4 4 3 2 2 3 • The new terminal spanning tree F is more expensive than the original terminal spanning tree T by the cycle-rule – p. 17/29

  32. The Bridge Lemma Lem (Bridge Lemma) For any terminal spanning tree T and any feasible fractional solution x to DCR, � C ∈C x C · br T ( C ) ≥ c ( T ) • Summarizing � � x C · br T ( C ) = x C · w ( Y C ) ≥ w ( F ) ≥ c ( T ) � �� � C ∈C C ∈C w -cost of � �� � integral w -cost of terminal fractional spanning tree terminal spanning tree – p. 18/29

  33. Approximation Factor – p. 19/29

  34. A First Bound Thr Algorithm IRR computes a solution of expected cost ≤ (1 + ln 2 + ε ) opt f – p. 20/29

  35. A First Bound Thr Algorithm IRR computes a solution of expected cost ≤ (1 + ln 2 + ε ) opt f Cor The integrality gap of DCR is at most 1 + ln 2 < 1 . 7 – p. 20/29

  36. A First Bound Thr Algorithm IRR computes a solution of expected cost ≤ (1 + ln 2 + ε ) opt f x t M c ( C )] ≤ 1 + ε X X X X E [ c ( C t )] ≤ C E [ opt f,t ] E [ apx ] = E [ M C t ≥ 1 t ≥ 1 t ≥ 1 M ln 2 ≤ 1 + ε opt f + 1 + ε X X E [ c ( T t )] M M t =1 t>M ln 2 • T t is a minimum terminal spanning tree at step t – p. 21/29

  37. A First Bound Thr Algorithm IRR computes a solution of expected cost ≤ (1 + ln 2 + ε ) opt f x t M c ( C )] ≤ 1 + ε X X X X E [ c ( C t )] ≤ C E [ opt f,t ] E [ apx ] = E [ M C t ≥ 1 t ≥ 1 t ≥ 1 M ln 2 ≤ 1 + ε opt f + 1 + ε X X E [ c ( T t )] M M t =1 t>M ln 2 Lem For any t , E [ c ( T t +1 )] ≤ (1 − 1 M ) c ( T t ) x t E [ c ( T t +1 )] ≤ c ( T t ) − E [ br T t ( C t )] = c ( T t ) − X C M br T t ( C ) C Bridge Lem c ( T t ) − 1 M c ( T t ) ≤ – p. 21/29

  38. A First Bound Thr Algorithm IRR computes a solution of expected cost ≤ (1 + ln 2 + ε ) opt f x t M c ( C )] ≤ 1 + ε X X X X E [ c ( C t )] ≤ C E [ opt f,t ] E [ apx ] = E [ M C t ≥ 1 t ≥ 1 t ≥ 1 M ln 2 ≤ 1 + ε opt f + 1 + ε X X E [ c ( T t )] M M t =1 t>M ln 2 Lem For any t , E [ c ( T t +1 )] ≤ (1 − 1 M ) c ( T t ) x t E [ c ( T t +1 )] ≤ c ( T t ) − E [ br T t ( C t )] = c ( T t ) − X C M br T t ( C ) C Bridge Lem c ( T t ) − 1 M c ( T t ) ≤ Cor E [ c ( T t )] ≤ (1 − 1 M ) t − 1 c ( T 1 ) ≤ (1 − 1 M ) t − 1 2 opt f – p. 21/29

  39. A First Bound Thr Algorithm IRR computes a solution of expected cost ≤ (1 + ln 2 + ε ) opt f x t M c ( C )] ≤ 1 + ε X X X X E [ c ( C t )] ≤ C E [ opt f,t ] E [ apx ] = E [ M C t ≥ 1 t ≥ 1 t ≥ 1 M ln 2 ≤ 1 + ε opt f + 1 + ε X X E [ c ( T t )] M M t =1 t>M ln 2 « t − 1 „ 1 1 − 1 ≤ opt f (1 + ε ) ln 2 + 2 opt f (1 + ε ) X M M t>M ln 2 ≤ (1 + ε )(ln 2 + 2 e − ln 2 ) · opt f – p. 21/29

  40. A Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (1 . 5 + ε ) opt – p. 22/29

  41. A Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (1 . 5 + ε ) opt Rem This bound might not hold w.r.t. opt f – p. 22/29

  42. A Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (1 . 5 + ε ) opt x t M c ( C )] ≤ 1 + ε X E [ c ( C t )] ≤ X X X E [ opt f,t ] C E [ apx ] = E [ M t ≥ 1 t ≥ 1 C t ≥ 1 M ln 4 ≤ 1 + ε E [ c ( S t )] + 1 + ε X X E [ c ( T t )] M M t =1 t> M ln 4 • S t is a minimum Steiner tree at step t – p. 22/29

  43. A Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (1 . 5 + ε ) opt Lem For any t , E [ c ( S t +1 )] ≤ (1 − 1 2 M ) c ( S t ) – p. 22/29

  44. A Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (1 . 5 + ε ) opt Lem For any t , E [ c ( S t +1 )] ≤ (1 − 1 2 M ) c ( S t ) • Construct a terminal spanning tree ( Y t , w ) w.r.t. S t and all its terminals R t = R ∩ S t as in the proof of the bridge lemma. – p. 22/29

  45. A Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (1 . 5 + ε ) opt Lem For any t , E [ c ( S t +1 )] ≤ (1 − 1 2 M ) c ( S t ) • Construct a terminal spanning tree ( Y t , w ) w.r.t. S t and all its terminals R t = R ∩ S t as in the proof of the bridge lemma. • Let b ( e ) ∈ S t be the bridge associated to e ∈ Y t . – p. 22/29

  46. A Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (1 . 5 + ε ) opt Lem For any t , E [ c ( S t +1 )] ≤ (1 − 1 2 M ) c ( S t ) • Construct a terminal spanning tree ( Y t , w ) w.r.t. S t and all its terminals R t = R ∩ S t as in the proof of the bridge lemma. • Let b ( e ) ∈ S t be the bridge associated to e ∈ Y t . b ( e 2 ) 4 e 2 4 b ( e 1 ) e 1 3 2 3 – p. 22/29

  47. A Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (1 . 5 + ε ) opt Lem For any t , E [ c ( S t +1 )] ≤ (1 − 1 2 M ) c ( S t ) • S ′ := S t - { b ( e ) ∈ S t | e ∈ br Y t , w ( C t ) } is a feasible Steiner tree at step t + 1 b ( e 2 ) 4 e 2 4 b ( e 1 ) e 1 3 2 3 – p. 23/29

  48. A Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (1 . 5 + ε ) opt Lem For any t , E [ c ( S t +1 )] ≤ (1 − 1 2 M ) c ( S t ) • S ′ := S t - { b ( e ) ∈ S t | e ∈ br Y t , w ( C t ) } is a feasible Steiner tree at step t + 1 b ( e 2 ) 4 e 2 4 b ( e 1 ) e 1 3 2 3 – p. 23/29

  49. A Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (1 . 5 + ε ) opt Lem For any t , E [ c ( S t +1 )] ≤ (1 − 1 2 M ) c ( S t ) • S ′ := S t - { b ( e ) ∈ S t | e ∈ br Y t , w ( C t ) } is a feasible Steiner tree at step t + 1 b ( e 2 ) 4 e 2 4 b ( e 1 ) 3 2 – p. 23/29

  50. A Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (1 . 5 + ε ) opt Lem For any t , E [ c ( S t +1 )] ≤ (1 − 1 2 M ) c ( S t ) • S ′ := S t - { b ( e ) ∈ S t | e ∈ br Y t , w ( C t ) } is a feasible Steiner tree at step t + 1 b ( e 2 ) 4 e 2 4 2 – p. 23/29

  51. A Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (1 . 5 + ε ) opt Lem For any t , E [ c ( S t +1 )] ≤ (1 − 1 2 M ) c ( S t ) E [ c ( S t +1 )] ≤ E [ c ( S ′ )] = c ( S t ) − E [ c ( { b ( e ) ∈ S t | e ∈ br Y t , w ( C t ) } )] = c ( S t ) − E [ br Y t , w ( C t )] = c ( S t ) − 1 X x t C br Y t , w ( C ) M C c ( S t ) − 1 Bridge Lem M w ( Y t ) ≤ = c ( S t ) − 1 M br S t , c ( R t ) c ( S t ) ≤ c ( S t ) − 1 2 M – p. 24/29

  52. A Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (1 . 5 + ε ) opt Lem For any t , E [ c ( S t +1 )] ≤ (1 − 1 2 M ) c ( S t ) E [ c ( S t +1 )] ≤ E [ c ( S ′ )] = c ( S t ) − E [ c ( { b ( e ) ∈ S t | e ∈ br Y t , w ( C t ) } )] = c ( S t ) − E [ br Y t , w ( C t )] = c ( S t ) − 1 X x t C br Y t , w ( C ) M C c ( S t ) − 1 Bridge Lem M w ( Y t ) ≤ = c ( S t ) − 1 M br S t , c ( R t ) c ( S t ) ≤ c ( S t ) − 1 2 M Cor E [ c ( S t )] ≤ (1 − 1 1 2 M ) t − 1 c ( S 1 ) = (1 − 2 M ) t − 1 opt – p. 24/29

  53. A Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (1 . 5 + ε ) opt x t j )] ≤ 1 + ε j X E [ c ( C t )] ≤ X X M c ( C t X E [ opt f,t ] E [ apx ] = E [ M j t ≥ 1 t ≥ 1 t ≥ 1 M ln 4 ≤ 1 + ε E [ c ( S t )] + 1 + ε X X E [ c ( T t )] M M t =1 t>M ln 4 M ln 4 ≤ (1 + ε 1 2(1 − 1 2 M ) t − 1 + X X M ) t − 1 ) opt ) · ( (1 − M t =1 t>M ln 4 ≤ (1 + ε )(2 − 2 e − ln(4) / 2 + 2 e − ln(4) ) · opt – p. 25/29

  54. An Even Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (ln 4 + ε ) opt – p. 26/29

  55. An Even Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (ln 4 + ε ) opt 1 5 2 4 3 1 • We define a random terminal spanning tree W ( witness tree) – p. 26/29

  56. An Even Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (ln 4 + ε ) opt 1 5 2 4 3 1 0 0 0 0 • We define a random terminal spanning tree W ( witness tree) – p. 26/29

  57. An Even Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (ln 4 + ε ) opt 1 5 2 4 3 1 • We define a random terminal spanning tree W ( witness tree) – p. 26/29

  58. An Even Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (ln 4 + ε ) opt 1 5 2 4 3 1 • We associate to each e in the Steiner tree S the edges W ( e ) of W such that the corresponding path in S contains e • Observe that | W ( e ) | is 1 , 2 . . . with probability 1 2 , 1 4 , . . . – p. 26/29

  59. An Even Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (ln 4 + ε ) opt 1 5 2 4 3 1 • We associate to each e in the Steiner tree S the edges W ( e ) of W such that the corresponding path in S contains e • Observe that | W ( e ) | is 1 , 2 . . . with probability 1 2 , 1 4 , . . . – p. 26/29

  60. An Even Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (ln 4 + ε ) opt 1 5 2 4 3 1 • For any sampled component C t , we delete from W a random set of bridges such that each edge of W is deleted with probability ≥ 1 /M ( ⇐ Farkas’ lemma+Bridge lemma) • When W ( e ) is deleted, we delete e from S – p. 26/29

  61. An Even Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (ln 4 + ε ) opt 1 5 2 4 3 1 • For any sampled component C t , we delete from W a random set of bridges such that each edge of W is deleted with probability ≥ 1 /M ( ⇐ Farkas’ lemma+Bridge lemma) • When W ( e ) is deleted, we delete e from S – p. 26/29

  62. An Even Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (ln 4 + ε ) opt 1 5 2 4 3 1 • For any sampled component C t , we delete from W a random set of bridges such that each edge of W is deleted with probability ≥ 1 /M ( ⇐ Farkas’ lemma+Bridge lemma) • When W ( e ) is deleted, we delete e from S – p. 26/29

  63. An Even Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (ln 4 + ε ) opt 1 5 2 4 3 1 • For any sampled component C t , we delete from W a random set of bridges such that each edge of W is deleted with probability ≥ 1 /M ( ⇐ Farkas’ lemma+Bridge lemma) • When W ( e ) is deleted, we delete e from S – p. 26/29

  64. An Even Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (ln 4 + ε ) opt 1 5 2 4 3 1 • For any sampled component C t , we delete from W a random set of bridges such that each edge of W is deleted with probability ≥ 1 /M ( ⇐ Farkas’ lemma+Bridge lemma) • When W ( e ) is deleted, we delete e from S – p. 26/29

  65. An Even Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (ln 4 + ε ) opt 1 5 2 3 1 • For any sampled component C t , we delete from W a random set of bridges such that each edge of W is deleted with probability ≥ 1 /M ( ⇐ Farkas’ lemma+Bridge lemma) • When W ( e ) is deleted, we delete e from S – p. 26/29

  66. An Even Better Bound Thr Algorithm IRR computes a solution of expected cost ≤ (ln 4 + ε ) opt 1 5 2 3 1 • Each e ∈ S survives in expectation M · ln 4 rounds – p. 26/29

  67. Derandomization Thr There is a ln 4 + ε deterministic approximation algorithm for Steiner tree – p. 27/29

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

Breaching the 2 -Approximation Barrier for Connectivity Augmentation: a Reduction to Steiner Tree

Breaching the 2 -Approximation Barrier for Connectivity Augmentation: a Reduction to Steiner Tree

Breaching the 2 -Approximation Barrier for Connectivity Augmentation: a Reduction to Steiner Tree Jarosaw Byrka 1 , Fabrizio Grandoni 2 , and Afrouz Jabal Ameli 2 1 University of Wrocaw 2 IDSIA, Lugano Abstract The basic goal of

251 views • 22 slides
On the integrality gap of hypergraphic Steiner tree relaxations Neil Olver Department of

On the integrality gap of hypergraphic Steiner tree relaxations Neil Olver Department of

On the integrality gap of hypergraphic Steiner tree relaxations Neil Olver Department of Mathematics, MIT Aussois, January 2012 Joint work with Michel Goemans, Thomas Rothvo and Rico Zenklusen. Steiner tree T erminals Steiner nodes 2 /

1.21k views • 51 slides
Pruned Dynamic Programming for Steiner Tree Yoichi Iwata (NII) Takuto Shigemura (U-Tokyo)

Pruned Dynamic Programming for Steiner Tree Yoichi Iwata (NII) Takuto Shigemura (U-Tokyo)

Separator-based Pruned Dynamic Programming for Steiner Tree Yoichi Iwata (NII) Takuto Shigemura (U-Tokyo) appeared in AAAI 2019 https://www.aaai.org/Papers/AAAI/2019/AAAI-IwataY.4578.pdf 1 Steiner Tree Problem Steiner tree problem Input:

755 views • 32 slides
Parameterized Approximation Schemes for Steiner Trees with Small Number of Steiner Vertices ak 1 ,

Parameterized Approximation Schemes for Steiner Trees with Small Number of Steiner Vertices ak 1 ,

Parameterized Approximation Schemes for Steiner Trees with Small Number of Steiner Vertices ak 1 , Andreas Emil Feldmann 1 , Du san Knop 1 , 2 ,Tom Pavel Dvo r a s k 1 , Tom s Toufar 1 , Pavel Vesel y 1 Masa r a 1

952 views • 40 slides
Approximation and Min-Max Results for the Steiner Connectivity Problem Marika Karbstein joint

Approximation and Min-Max Results for the Steiner Connectivity Problem Marika Karbstein joint

Approximation and Min-Max Results for the Steiner Connectivity Problem Marika Karbstein joint work with Ralf Borndrfer Zuse Institute Berlin DFG Research Center M ATHEON Mathematics for key technologies Aussois, January 2014 The Steiner

1.07k views • 47 slides
FOARS: FLUTE Based Obstacle- Avoiding Rectilinear Steiner Tree Construction Gaurav Ajwani and

FOARS: FLUTE Based Obstacle- Avoiding Rectilinear Steiner Tree Construction Gaurav Ajwani and

FOARS: FLUTE Based Obstacle- Avoiding Rectilinear Steiner Tree Construction Gaurav Ajwani and Chris Chu Iowa State University Wai-Kei Mak National Tsing Hua University 1 OARSMT Problem Formulation Obstacle-avoiding rectilinear Steiner

459 views • 25 slides
Critical-Trunk Based Obstacle-Avoiding Rectilinear Steiner Tree Routings for Delay and Slack

Critical-Trunk Based Obstacle-Avoiding Rectilinear Steiner Tree Routings for Delay and Slack

Critical-Trunk Based Obstacle-Avoiding Rectilinear Steiner Tree Routings for Delay and Slack Optimization ISPD-2009 Yen-Hung Lin, Shu-Hsin Chang, Yih-Lang Li Dept. of CSIE, National Chiao Tung University Institute of Computer Science and

482 views • 22 slides
Efficient Multilayer Routing Based on Obstacle-Avoiding Preferred Direction Steiner Tree

Efficient Multilayer Routing Based on Obstacle-Avoiding Preferred Direction Steiner Tree

Efficient Multilayer Routing Based on Obstacle-Avoiding Preferred Direction Steiner Tree Ching-Hung Liu, Yao-Hsin Chou, Shin-Yi Yuan, and Sy-Yen Kuo National Taiwan University 1 2 Outline 3 Outline The RSMT Problem Given a set of pins,

622 views • 37 slides
Directed Steiner Tree and the Lasserre Hierarchy Thomas Rothvo Department of Mathematics,

Directed Steiner Tree and the Lasserre Hierarchy Thomas Rothvo Department of Mathematics,

Directed Steiner Tree and the Lasserre Hierarchy Thomas Rothvo Department of Mathematics, M.I.T. Directed Steiner Tree r 4 7 9 8 6 4 8 7 Directed Steiner Tree Input: directed weighted graph G = ( V, E, c ) r 3 2 4 7 3 5 2

1.08k views • 88 slides
Steiner-Star-Free Graphs and Equivalence of Steiner Tree Relaxations Andreas Emil Feldmann 1

Steiner-Star-Free Graphs and Equivalence of Steiner Tree Relaxations Andreas Emil Feldmann 1

Steiner-Star-Free Graphs and Equivalence of Steiner Tree Relaxations Andreas Emil Feldmann 1 Jochen Knemann 1 Neil Olver 2 Laura Sanit 1 1 Combinatorics &amp; Optimization, University of Waterloo 2 VU University &amp; CWI, Amsterdam The

1.18k views • 41 slides
A Note on Online Steiner Tree Problems Gokarna Sharma and Costas Busch Division of Computer

A Note on Online Steiner Tree Problems Gokarna Sharma and Costas Busch Division of Computer

A Note on Online Steiner Tree Problems Gokarna Sharma and Costas Busch Division of Computer Science and Eng. Louisiana State University CCCG 2014 1 Steiner tree problem Given an arbitrary undirected graph with weights G (lengths) on

618 views • 43 slides
Steiner Forest (Approximation Algorithms by V. Vazirani, Chapter 22) - Problem statement;

Steiner Forest (Approximation Algorithms by V. Vazirani, Chapter 22) - Problem statement;

Steiner Forest (Approximation Algorithms by V. Vazirani, Chapter 22) - Problem statement; integer program for Steiner Forest, LP-relaxation and its dual program; comparison with the Steiner Network problem - Primal-dual schema, primal

133 views • 11 slides
Tetraquarks in the Steiner tree model of confinement available at

Tetraquarks in the Steiner tree model of confinement available at

Introduction Symmetry breaking and tetraquarks The additive model of tetraquark confinement The Steiner-tree model of confinement Conclusions Tetraquarks in the Steiner tree model of confinement available at

644 views • 20 slides
ug[SCIP-Jack, MPI]: A Massively Parallel Steiner Tree Solver Daniel Rehfeldt Yuji Shinano

ug[SCIP-Jack, MPI]: A Massively Parallel Steiner Tree Solver Daniel Rehfeldt Yuji Shinano

ug[SCIP-Jack, MPI]: A Massively Parallel Steiner Tree Solver Daniel Rehfeldt Yuji Shinano Thorsten Koch Zuse Institute Berlin Technische Universit at Berlin UG19, Berlin, January 2019 The Steiner Tree Problem in Graphs Given: G =

757 views • 60 slides
Approximation Algorithms in low-dimensional geometry or on Planar Graphs Claire Mathieu

Approximation Algorithms in low-dimensional geometry or on Planar Graphs Claire Mathieu

Approximation Algorithms in low-dimensional geometry or on Planar Graphs Claire Mathieu Thanks to Klein and Borradaile for many slides Joint work with Borradaile and Klein Steiner tree Terminal Aroras geometric PTAS technique: Break

895 views • 46 slides
Wot the L: Analysis of Real versus Random Placed Nets, and Implications for Steiner Tree

Wot the L: Analysis of Real versus Random Placed Nets, and Implications for Steiner Tree

Wot the L: Analysis of Real versus Random Placed Nets, and Implications for Steiner Tree Heuristics Andrew B. Kahng, Christopher Moyes, Sriram Venkatesh and Lutong Wang UC San Diego CSE and ECE Departments Outline Background and

427 views • 23 slides
Cartoon parallel architectures; CPUs and GPUs CSE 6230, Fall 2014 Th Sep 11 Thanks

Cartoon parallel architectures; CPUs and GPUs CSE 6230, Fall 2014 Th Sep 11 Thanks

Cartoon parallel architectures; CPUs and GPUs CSE 6230, Fall 2014 Th Sep 11 Thanks to Jee Choi (a senior PhD student) for a big assist 1 2 3 4 5 6 7 8 9 10 11 12 13 14 ~ socket 14 ~ core 14 ~ HWMT+SIMD

1.02k views • 72 slides
Milestones in Descriptive Complexity Theory ANU Logic Summer School, 2019 Presented by Sasha

Milestones in Descriptive Complexity Theory ANU Logic Summer School, 2019 Presented by Sasha

Milestones in Descriptive Complexity Theory ANU Logic Summer School, 2019 Presented by Sasha Rubin School of Computer Science University of Sydney What is complexity theory? Analyses and classifies problems by the amount of a resource

1.36k views • 119 slides
Stronger Connections Between Circuit Analysis and Circuit Lower Bounds, via PCPs of Proximity

Stronger Connections Between Circuit Analysis and Circuit Lower Bounds, via PCPs of Proximity

Stronger Connections Between Circuit Analysis and Circuit Lower Bounds, via PCPs of Proximity Lijie Chen Ryan Williams Context: The Algorithmic Method for Proving Circuit Lower Bounds Proving limitations on non-uniform circuits is extremely

796 views • 31 slides
THE ATTRIBUTOR: A VERSATILE INTER-PROCEDURAL FIXPOINT ITERATION FRAMEWORK LLVM-Dev19

THE ATTRIBUTOR: A VERSATILE INTER-PROCEDURAL FIXPOINT ITERATION FRAMEWORK LLVM-Dev19

THE ATTRIBUTOR: A VERSATILE INTER-PROCEDURAL FIXPOINT ITERATION FRAMEWORK LLVM-Dev19 October 22, 2019 San Jose, CA, USA Johannes Doerfert*, Hideto Ueno, Stefan Stipanovic *Leadership Computing Facility *Argonne National Laboratory *

1.14k views • 92 slides
Zero forcing, propagation time, and throttling on a graph Leslie Hogben Iowa State University

Zero forcing, propagation time, and throttling on a graph Leslie Hogben Iowa State University

Zero forcing, propagation time, and throttling on a graph Leslie Hogben Iowa State University and American Institute of Mathematics New York Combinatorics Seminar August 28, 2020 Leslie Hogben (Iowa State University and American Institute of

752 views • 53 slides
Inference and Optimalities in Estimation of Gaussian Graphical Model Harrison H. Zhou Department

Inference and Optimalities in Estimation of Gaussian Graphical Model Harrison H. Zhou Department

Inference and Optimalities in Estimation of Gaussian Graphical Model Harrison H. Zhou Department of Statistics Yale University Jointly with Zhao Ren, Tingni Sun and Cun-Hui Zhang 1 Outline Introduction Main Results Asymptotic

509 views • 27 slides
Energy-Efficient Management of Virtual Machines in Data Centers for Cloud Computing Anton

Energy-Efficient Management of Virtual Machines in Data Centers for Cloud Computing Anton

H EURISTICS M ARKOV H OST O VERLOAD D ETECTION I MPLEMENTATION C ONCLUSIONS Energy-Efficient Management of Virtual Machines in Data Centers for Cloud Computing Anton Beloglazov Supervisor: Prof. Rajkumar Buyya The Cloud Computing and

995 views • 87 slides
The natural emergence of the SFR-H2 surface density relation in galaxy simulations Alessandro

The natural emergence of the SFR-H2 surface density relation in galaxy simulations Alessandro

The natural emergence of the SFR-H2 surface density relation in galaxy simulations Alessandro Lupi (Institut dAstrophysique de Paris) THE ROLE OF GAS IN GALAXY DYNAMICS with: S. Bovino, P. R. Capelo, M. Volonteri, J. Silk October 2nd, 2017

542 views • 14 slides

More recommend