distributed algorithms below the
play

Distributed Algorithms Below the Greedy Regime Yannic Maus Mohsen - PowerPoint PPT Presentation

Coloring Distributed Algorithms Below the Greedy Regime Yannic Maus Mohsen Ghaffari , Juho Hirvonen, Fabian Kuhn, Jara Uitto This project has received funding from the European Unions Horizon 2020 Research and Innovation Programme under


  1. Coloring Distributed Algorithms Below the Greedy Regime Yannic Maus Mohsen Ghaffari , Juho Hirvonen, Fabian Kuhn, Jara Uitto This project has received funding from the European Union’s Horizon 2020 Research and Innovation Programme under grant agreement no. 755839.

  2. LOCAL Model [Linial ; FOCS ’87] 𝑯 = 𝑾, 𝑭 , Communication Network = Problem Instance: 𝒐 = 𝑾 I am green 11 11 11 15 15 15 5 5 5 6 6 6 1 1 1 2 2 2 9 9 9 21 21 21 7 7 7 26 26 26 33 33 33 8 8 8 27 27 27 Discrete synchronous rounds: β€’ local computations β€’ exchange messages with all neighbors (computations unbounded, message sizes are unbounded) time complexity = number of rounds 3

  3. CONGEST MODEL 𝑯 = 𝑾, 𝑭 , Communication Network = Problem Instance: 𝒐 = 𝑾 11 11 15 15 5 5 6 6 1 1 2 2 9 9 21 21 7 7 26 26 33 33 8 8 27 27 Discrete synchronous rounds: β€’ local computations β€’ exchange messages with all neighbors 𝑷(𝐦𝐩𝐑 𝒐) bits (computations unbounded, message sizes are unbounded) ) time complexity = number of rounds 4

  4. Classic Big Four (Greedy Regime) 𝚬 + 𝟐 -Vertex Coloring Maximal Ind. Set (MIS) (πŸ‘πš¬ βˆ’ 𝟐) -Edge Coloring Maximal Matching ( Ξ” : maximum degree of 𝐻 ) 5

  5. In the LOCAL Model … Greedy Below Greedy 2 𝑃 log π‘œ 2 𝑃 log π‘œ Maximal IS Maximum IS, 𝟐 βˆ’ 𝝑 -approx. poly log π‘œ vertex cover, 2 𝑃 log π‘œ vertex cover, 𝟐 + 𝝑 -approx. 2 -approx. 2 𝑃 log π‘œ 2 𝑃 log π‘œ min. dominating set, min dominating set, ( 𝟐 + 𝝑) 𝐦𝐩𝐑 𝚬 -approx. 𝟐 + 𝝑 -approx. 2 𝑃 log π‘œ 2 𝑃 log π‘œ hypergraph vertex cover, hypergraph vertex cover, 𝟐 + 𝝑 -approx. rank -approx. 𝚬 + 𝟐 -vertex coloring 𝚬 -vertex coloring 2 𝑃 log π‘œ 2 𝑃 log π‘œ πŸ‘πš¬ βˆ’ 𝟐 -edge coloring poly log π‘œ poly log π‘œ 𝟐 + 𝝑 𝚬 -edge coloring poly log π‘œ poly log π‘œ maximal matching Maximum Matching, 𝟐 + 𝝑 -approx. CONGEST 6

  6. In the LOCAL Model … Greedy Below Greedy 2 𝑃 log π‘œ 2 𝑃 log π‘œ Maximal IS Maximum IS, 𝟐 βˆ’ 𝝑 -approx. β€œProblems that do have β€œ Problems that do not have easy sequential greedy easy sequential greedy algorithms .” algorithms .” 𝚬 + 𝟐 -vertex coloring 𝚬 -vertex coloring 2 𝑃 log π‘œ 2 𝑃 log π‘œ πŸ‘πš¬ βˆ’ 𝟐 -edge coloring poly log π‘œ poly log π‘œ 𝟐 + 𝝑 𝚬 -edge coloring poly log π‘œ poly log π‘œ maximal matching Maximum Matching, 𝟐 + 𝝑 -approx. CONGEST 7

  7. Outline This Talk: How do we use LOCAL? Below Greedy Ξ©(π‘œ 2 ) 2 𝑃 log π‘œ Maximum IS Technique 1: Ball growing Maximum IS, 𝟐 βˆ’ 𝝑 -approx. 7/8 -approx. [Ghaffari, Kuhn, Maus; STOC ’ 17 ] Ξ©(π‘œ 2 ) vertex cover, 2 𝑃 log π‘œ vertex cover, 𝟐 + 𝝑 -approx. exact Ξ©(π‘œ 2 ) 2 𝑃 log π‘œ min. dominating set, min dominating set, 𝟐 + 𝝑 -approx. exact Ξ©(π‘œ 2 ) 2 𝑃 log π‘œ hypergraph vertex cover, hypergraph vertex cover, 𝟐 + 𝝑 -approx. exact Ξ©(π‘œ 2 ) 𝝍(𝑯) -vertex coloring 𝚬 -vertex coloring Technique 2: Local filling 2 𝑃 log π‘œ [Ghaffari, Hirvonen, Kuhn, Maus; PODC ’ 18 ] ? poly log π‘œ 𝟐 + 𝝑 𝚬 -edge coloring edge coloring Technique 3: Aug. paths ? poly log π‘œ [Ghaffari, Kuhn, Maus , Uitto; STOC ’ 18 ] Maximum matching Maximum Matching, 𝟐 βˆ’ 𝝑 -approx. exact CONGEST 8

  8. Technique 1: Ball Growing This Talk: How do we use LOCAL? Below Greedy Ξ©(π‘œ 2 ) 2 𝑃 log π‘œ Maximum IS Technique 1: Ball growing Maximum IS, 𝟐 βˆ’ 𝝑 -approx. 7/8 -approx. Ξ©(π‘œ 2 ) vertex cover, 2 𝑃 log π‘œ vertex cover, 𝟐 + 𝝑 -approx. exact Ξ©(π‘œ 2 ) 2 𝑃 log π‘œ min. dominating set, min dominating set, 𝟐 + 𝝑 -approx. exact Ξ©(π‘œ 2 ) 2 𝑃 log π‘œ hypergraph vertex cover, hypergraph vertex cover, 𝟐 + 𝝑 -approx. exact Ξ©(π‘œ 2 ) 𝝍(𝑯) -vertex coloring 𝚬 -vertex coloring Technique 2: Local filling 2 𝑃 log π‘œ ? poly log π‘œ 𝟐 + 𝝑 𝚬 -edge coloring edge coloring Technique 3: Aug. paths ? poly log π‘œ Maximum matching Maximum Matching, 𝟐 βˆ’ 𝝑 -approx. exact 9

  9. Sequential Ball Growing (MaxIS) π’˜ Safe Ball π‘ͺ 𝒔 (π’˜) : |𝑁𝑏𝑦𝐽𝑇(𝐢 𝑠+1 )| < (1 + πœ—) β‹… 𝑁𝑏𝑦𝐽𝑇(𝐢 𝑠 ) Find safe ball: Set 𝑠 = 0 and increase 𝑠 until ball 𝐢 𝑠 is safe. Terminates with small radius 𝑠 = 𝑃(πœ— βˆ’1 log π‘œ) . 10

  10. Sequential Ball Growing (MaxIS) Safe Ball π‘ͺ 𝒔 (π’˜) : |𝑁𝑏𝑦𝐽𝑇(𝐢 𝑠+1 )| < (1 + πœ—) β‹… 𝑁𝑏𝑦𝐽𝑇(𝐢 𝑠 ) Find safe ball: Set 𝑠 = 0 and increase 𝑠 until ball 𝐢 𝑠 is safe. Terminates with small radius 𝑠 = 𝑃(πœ— βˆ’1 log π‘œ) . 11

  11. Sequential Ball Growing (MaxIS) Safe Ball π‘ͺ 𝒔 (π’˜) : |𝑁𝑏𝑦𝐽𝑇(𝐢 𝑠+1 )| < (1 + πœ—) β‹… 𝑁𝑏𝑦𝐽𝑇(𝐢 𝑠 ) Find safe ball: Set 𝑠 = 0 and increase 𝑠 until ball 𝐢 𝑠 is safe. Terminates with small radius 𝑠 = 𝑃(πœ— βˆ’1 log π‘œ) . 12

  12. Sequential Ball Growing (MaxIS) Safe Ball π‘ͺ 𝒔 (π’˜) : |𝑁𝑏𝑦𝐽𝑇(𝐢 𝑠+1 )| < (1 + πœ—) β‹… 𝑁𝑏𝑦𝐽𝑇(𝐢 𝑠 ) Find safe ball: Set 𝑠 = 0 and increase 𝑠 until ball 𝐢 𝑠 is safe. Terminates with small radius 𝑠 = 𝑃(πœ— βˆ’1 log π‘œ) . 13

  13. Sequential Ball Growing (MaxIS) Safe Ball π‘ͺ 𝒔 (π’˜) : |𝑁𝑏𝑦𝐽𝑇(𝐢 𝑠+1 )| < (1 + πœ—) β‹… 𝑁𝑏𝑦𝐽𝑇(𝐢 𝑠 ) Find safe ball: Set 𝑠 = 0 and increase 𝑠 until ball 𝐢 𝑠 is safe. Terminates with small radius 𝑠 = 𝑃(πœ— βˆ’1 log π‘œ) . 14

  14. Sequential Ball Growing (MaxIS) Sequentially computes a 1 + πœ— βˆ’1 -approximation for MaxIS. (using unbounded computation) Safe Ball π‘ͺ 𝒔 (π’˜) : |𝑁𝑏𝑦𝐽𝑇(𝐢 𝑠+1 )| < (1 + πœ—) β‹… 𝑁𝑏𝑦𝐽𝑇(𝐢 𝑠 ) Find safe ball: Set 𝑠 = 0 and increase 𝑠 until ball 𝐢 𝑠 is safe. Terminates with small radius 𝑠 = 𝑃(πœ— βˆ’1 log π‘œ) . 15

  15. Parallel Ball Growing Theorem Using ( πͺ𝐩𝐦𝐳 𝐦𝐩𝐑 𝒐 , πͺ𝐩𝐦𝐳 𝐦𝐩𝐑 𝒐) -network decompositions β€œsequentially ball growing” can be β€œdone in parallel” in LOCAL. [ STOC ’ 17, Ghaffari, Kuhn, Maus ] Corollary 𝐦𝐩𝐑 𝒐 deterministic There are πͺ𝐩𝐦𝐳 𝐦𝐩𝐑 𝒐 randomized and πŸ‘ 𝑷 𝟐 + 𝝑 -approximation algorithms for covering and packing integer linear programs . This includes maximum independent set, minimum dominating set, vertex cover, … . [ STOC ’ 17, Ghaffari, Kuhn, Maus ] 16

  16. Technique 2: Local Filling This Talk: How do we use LOCAL? Below Greedy Ξ©(π‘œ 2 ) 2 𝑃 log π‘œ Maximum IS Technique 1: Ball growing Maximum IS, 𝟐 βˆ’ 𝝑 -approx. 7/8 -approx. Ξ©(π‘œ 2 ) vertex cover, 2 𝑃 log π‘œ vertex cover, 𝟐 + 𝝑 -approx. exact Ξ©(π‘œ 2 ) 2 𝑃 log π‘œ min. dominating set, min dominating set, 𝟐 + 𝝑 -approx. exact Ξ©(π‘œ 2 ) 2 𝑃 log π‘œ hypergraph vertex cover, hypergraph vertex cover, 𝟐 + 𝝑 -approx. exact Ξ©(π‘œ 2 ) 𝝍(𝑯) -vertex coloring 𝚬 -vertex coloring Technique 2: Local filling 2 𝑃 log π‘œ ? poly log π‘œ 𝟐 + 𝝑 𝚬 -edge coloring edge coloring Technique 3: Aug. paths ? poly log π‘œ Maximum matching Maximum Matching, 𝟐 βˆ’ 𝝑 -approx. exact 17

  17. Ξ” -Coloring Previous Work: [Panconesi, Srinivasan; STOC ’93 ] Definition: An induced subgraph 𝐼 βŠ† 𝐻 is called an easy component if any 𝛦 𝐻 -coloring of 𝐻 βˆ– 𝐼 can be extended to a 𝛦 𝐻 -coloring of 𝐻 without changing the coloring on 𝐻 βˆ– 𝐼 . Well studied under the name degree chosable components. [ ErdΕ‘s et al. ’79, Vizing β€˜76 ] β€œ β€œ Theorem: Let 𝐻 be a graph ( β‰  clique) with max. degree Ξ” β‰₯ 3 . Every node of 𝐻 has a small diameter easy component in distance at most O(log n) . [ PODC ’ 18; Ghaffari, Hirvonen, Kuhn, Maus ] 18

  18. Find an MIS 𝑡 of small diameter easy components Define 𝑃 log π‘œ Layers: 𝑴 𝒋 = 𝑀 𝑀 in distance 𝒋 to some component in 𝑡} For 𝑗 = 𝑃(log π‘œ) to 1 color nodes in 𝑀 𝑗 through solving a (deg+1)-list coloring Color easy components in M 19

  19. Find an MIS 𝑡 of small diameter easy components Define 𝑃 log π‘œ Layers: 𝑴 𝒋 = 𝑀 𝑀 in distance 𝒋 to some component in 𝑡} For 𝑗 = 𝑃(log π‘œ) to 1 color nodes in 𝑀 𝑗 through solving a (deg+1)-list coloring Color easy components in M 20

  20. Find an MIS 𝑡 of small diameter easy components Greedy regime Define 𝑃 log π‘œ Layers: 𝑴 𝒋 = 𝑀 𝑀 in distance 𝒋 to some component in 𝑡} For 𝑗 = 𝑃(log π‘œ) to 1 color nodes in 𝑀 𝑗 through solving a (deg+1)-list coloring Color easy components in M 21

  21. Find an MIS 𝑡 of small diameter easy components Greedy regime Define 𝑃 log π‘œ Layers: 𝑴 𝒋 = 𝑀 𝑀 in distance 𝒋 to some component in 𝑡} For 𝑗 = 𝑃(log π‘œ) to 1 color nodes in 𝑀 𝑗 through solving a (deg+1)-list coloring Color easy components in M 22

  22. Find an MIS 𝑡 of small diameter easy components Greedy regime Define 𝑃 log π‘œ Layers: 𝑴 𝒋 = 𝑀 𝑀 in distance 𝒋 to some component in 𝑡} For 𝑗 = 𝑃(log π‘œ) to 1 color nodes in 𝑀 𝑗 through solving a (deg+1)-list coloring Color easy components in M 23

Recommend


More recommend