honors combinatorics
play

Honors Combinatorics CMSC-27410 = Math-28410 CMSC-37200 Instructor: - PowerPoint PPT Presentation

Honors Combinatorics CMSC-27410 = Math-28410 CMSC-37200 Instructor: Laszlo Babai University of Chicago Week 5, Thursday, May 7, 2020 CMSC-27410=Math-28410 CMSC-3720 Honors Combinatorics Fractional cover hypergraph H = ( V , E ) V = { v


  1. Honors Combinatorics CMSC-27410 = Math-28410 ∼ CMSC-37200 Instructor: Laszlo Babai University of Chicago Week 5, Thursday, May 7, 2020 CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  2. Fractional cover hypergraph H = ( V , E ) V = { v 1 , . . . , v n } E = { E 1 , . . . , E m } fractional cover y = ( y 1 , . . . , y n ) ∈ R n putting weight y i on vertex v i , s.t. ( ∀ i ∈ [ n ])( y i ≥ 0 ) (: weights non-negative :) ( ∀ j ∈ [ m ])( � i : v i ∈ E j y i ≥ 1 ) (: total weight on each edge ≥ 1 :) τ ∗ ( H ) = min { � n i = 1 y i | constraints } τ ( H ) ≥ τ ∗ ( H ) Integral optimum: CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  3. Fractional matching hypergraph H = ( V , E ) V = { v 1 , . . . , v n } E = { E 1 , . . . , E m } fractional matching x = ( x 1 , . . . , x m ) ∈ R m putting weight x j on edge E j , s.t. ( ∀ j ∈ [ m ])( x j ≥ 0 ) (: weights non-negative :) ( ∀ i ∈ [ n ])( � j : v i ∈ E j x j ≤ 1 ) (: total weight on each vertex ≤ 1 :) ν ∗ ( H ) = max { � m j = 1 x j | constraints } ν ( H ) ≤ ν ∗ ( H ) = τ ∗ ( H ) Integral optimum: CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  4. Greedy cover algorithm hypergraph H = ( V , E ) V = { v 1 , . . . , v n } , E = { E 1 , . . . , E m } T ← ∅ while E � ∅ do v ← argmax { deg ( v ) | v ∈ V } (: max degree :) E ← E \ { edges incident with v } T ← T ∪ { x } end ( while ) τ greedy := | T | (: size of greedy cover :) CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  5. Fractional vs. greedy cover size of greedy cover τ greedy ν ≤ ν ∗ = τ ∗ ≤ τ ≤ τ greedy CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  6. Fractional vs. greedy cover size of greedy cover τ greedy ν ≤ ν ∗ = τ ∗ ≤ τ ≤ τ greedy Theorem (Lovász 1975) τ greedy < ( 1 + ln deg max ) · τ ∗ Corollary: Integrality gap τ/τ ∗ < 1 + ln deg max Corollary: Approximation ratio τ greedy /τ < 1 + ln deg max CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  7. Fractional vs. greedy cover: proof k -matching: set of edges s.t. ∀ vertex covered ≤ k times ν k max size of k -matching ν = ν 1 i.e. subhypergraph with deg max ≤ k CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  8. Fractional vs. greedy cover: proof k -matching: set of edges s.t. ∀ vertex covered ≤ k times ν k max size of k -matching ν = ν 1 i.e. subhypergraph with deg max ≤ k ν ∗ ≥ ν k / k CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  9. Fractional vs. greedy cover: proof k -matching: set of edges s.t. ∀ vertex covered ≤ k times ν k max size of k -matching ν = ν 1 i.e. subhypergraph with deg max ≤ k ν ∗ ≥ ν k / k why? CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  10. Fractional vs. greedy cover: proof k -matching: set of edges s.t. ∀ vertex covered ≤ k times ν k max size of k -matching ν = ν 1 i.e. subhypergraph with deg max ≤ k ν ∗ ≥ ν k / k why? put weight 1 / k on each edge of k -matching total weight on ∀ vertex ≤ 1 CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  11. Fractional vs. greedy cover: proof k -matching: set of edges s.t. ∀ vertex covered ≤ k times ν k max size of k -matching ν = ν 1 i.e. subhypergraph with deg max ≤ k ν ∗ ≥ ν k / k why? put weight 1 / k on each edge of k -matching total weight on ∀ vertex ≤ 1 Max degree starts at d := deg max then goes down t i := # rounds while max degree = i τ greedy = t d + t d − 1 + · · · + t 1 (total # rounds) CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  12. Fractional vs. greedy cover: proof t i := # rounds while max degree = i τ greedy = t d + t d − 1 + · · · + t 1 (total # rounds) E i set of edges left over after t d + t d − 1 + · · · + t i + 1 rounds CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  13. Fractional vs. greedy cover: proof t i := # rounds while max degree = i τ greedy = t d + t d − 1 + · · · + t 1 (total # rounds) E i set of edges left over after t d + t d − 1 + · · · + t i + 1 rounds E i has deg max ≤ i for the first time starts sequence of t i rounds with deg max = i CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  14. Fractional vs. greedy cover: proof t i := # rounds while max degree = i τ greedy = t d + t d − 1 + · · · + t 1 (total # rounds) E i set of edges left over after t d + t d − 1 + · · · + t i + 1 rounds E i has deg max ≤ i for the first time starts sequence of t i rounds with deg max = i ( t i = 0 possible; then E i = E i − 1 ) CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  15. Fractional vs. greedy cover: proof t i := # rounds while max degree = i τ greedy = t d + t d − 1 + · · · + t 1 (total # rounds) E i set of edges left over after t d + t d − 1 + · · · + t i + 1 rounds E i has deg max ≤ i for the first time starts sequence of t i rounds with deg max = i ( t i = 0 possible; then E i = E i − 1 ) |E i | = i · t i + ( i − 1 ) · t i − 1 + · · · + 1 · t 1 CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  16. Fractional vs. greedy cover: proof t i := # rounds while max degree = i τ greedy = t d + t d − 1 + · · · + t 1 (total # rounds) E i set of edges left over after t d + t d − 1 + · · · + t i + 1 rounds CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  17. Fractional vs. greedy cover: proof t i := # rounds while max degree = i τ greedy = t d + t d − 1 + · · · + t 1 (total # rounds) E i set of edges left over after t d + t d − 1 + · · · + t i + 1 rounds |E i | = i · t i + ( i − 1 ) · t i − 1 + · · · + 1 · t 1 CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  18. Fractional vs. greedy cover: proof t i := # rounds while max degree = i τ greedy = t d + t d − 1 + · · · + t 1 (total # rounds) E i set of edges left over after t d + t d − 1 + · · · + t i + 1 rounds |E i | = i · t i + ( i − 1 ) · t i − 1 + · · · + 1 · t 1 E i i -matching: deg max ≤ i |E i | ≤ ν i ≤ i · ν ∗ ∴ CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  19. Fractional vs. greedy cover: proof t i := # rounds while max degree = i τ greedy = t d + t d − 1 + · · · + t 1 (total # rounds) E i set of edges left over after t d + t d − 1 + · · · + t i + 1 rounds |E i | = i · t i + ( i − 1 ) · t i − 1 + · · · + 1 · t 1 E i i -matching: deg max ≤ i |E i | ≤ ν i ≤ i · ν ∗ ∴ i · t i + ( i − 1 ) · t i − 1 + · · · + 1 · t 1 ≤ i · ν ∗ CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  20. Fractional vs. greedy cover: proof t i := # rounds while max degree = i τ greedy = t d + t d − 1 + · · · + t 1 (total # rounds) L i := i · t i + ( i − 1 ) · t i − 1 + · · · + 1 · t 1 ≤ i · ν ∗ CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  21. Fractional vs. greedy cover: proof t i := # rounds while max degree = i τ greedy = t d + t d − 1 + · · · + t 1 (total # rounds) L i := i · t i + ( i − 1 ) · t i − 1 + · · · + 1 · t 1 ≤ i · ν ∗ 1 1 ( d − 1 ) · d L d − 1 + 1 1 K := 1 · 2 L 1 + 2 · 3 L 2 + · · · + d L d CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  22. Fractional vs. greedy cover: proof t i := # rounds while max degree = i τ greedy = t d + t d − 1 + · · · + t 1 (total # rounds) L i := i · t i + ( i − 1 ) · t i − 1 + · · · + 1 · t 1 ≤ i · ν ∗ 1 1 ( d − 1 ) · d L d − 1 + 1 1 K := 1 · 2 L 1 + 2 · 3 L 2 + · · · + d L d coeff ( t i ) = � � 1 1 ( d − 1 ) d + 1 1 i i ( i + 1 ) + ( i + 1 )( i + 2 ) + · · · + d CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  23. A telescoping sum 1 1 ( d − 1 ) d + 1 1 i ( i + 1 ) + ( i + 1 )( i + 2 ) + · · · + d CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  24. A telescoping sum 1 1 ( d − 1 ) d + 1 1 i ( i + 1 ) + ( i + 1 )( i + 2 ) + · · · + d � 1 1 1 1 d − 1 − 1 1 + 1 � � � � � = i − + i + 1 − + · · · + i + 1 i + 2 d d = 1 i CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  25. Fractional vs. greedy cover: proof τ greedy = t d + t d − 1 + · · · + t 1 (total # rounds) L i := i · t i + ( i − 1 ) · t i − 1 + · · · + 1 · t 1 ≤ i · ν ∗ 1 1 ( d − 1 ) · d L d − 1 + 1 1 2 · 3 L 2 + · · · + K := 1 · 2 L 1 + d L d coeff ( t i ) = � � 1 1 ( d − 1 ) d + 1 1 i i ( i + 1 ) + ( i + 1 )( i + 2 ) + · · · + d CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  26. Fractional vs. greedy cover: proof τ greedy = t d + t d − 1 + · · · + t 1 (total # rounds) L i := i · t i + ( i − 1 ) · t i − 1 + · · · + 1 · t 1 ≤ i · ν ∗ 1 1 ( d − 1 ) · d L d − 1 + 1 1 K := 1 · 2 L 1 + 2 · 3 L 2 + · · · + d L d coeff ( t i ) = � � 1 1 ( d − 1 ) d + 1 1 = 1 i i ( i + 1 ) + ( i + 1 )( i + 2 ) + · · · + d K = t d + t d − 1 + · · · + d 1 = CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

  27. Fractional vs. greedy cover: proof τ greedy = t d + t d − 1 + · · · + t 1 (total # rounds) L i := i · t i + ( i − 1 ) · t i − 1 + · · · + 1 · t 1 ≤ i · ν ∗ 1 1 ( d − 1 ) · d L d − 1 + 1 1 K := 1 · 2 L 1 + 2 · 3 L 2 + · · · + d L d coeff ( t i ) = � � 1 1 ( d − 1 ) d + 1 1 = 1 i i ( i + 1 ) + ( i + 1 )( i + 2 ) + · · · + d K = t d + t d − 1 + · · · + d 1 = τ greedy CMSC-27410=Math-28410 ∼ CMSC-3720 Honors Combinatorics

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