13
Christoph Lenzen, PODC 2011 Christoph Lenzen, PODC 2011 MIS on Trees Christoph Lenzen and Roger Wattenhofer

MIS on Trees Christoph Lenzen and Roger Wattenhofer

  • Upload
    farren

  • View
    38

  • Download
    0

Embed Size (px)

DESCRIPTION

MIS on Trees Christoph Lenzen and Roger Wattenhofer. What is a Maximal Independet Set (MIS)?. inaugmentable set of non-adjacent nodes well-known symmetry breaking structure many algorithms build on a MIS. What is a Tree?. Let’s assume we all know. Talk Outline. in each phase: - PowerPoint PPT Presentation

Citation preview

Page 1: MIS on Trees Christoph Lenzen and Roger Wattenhofer

Christoph Lenzen, PODC 2011Christoph Lenzen, PODC 2011

MIS on TreesChristoph Lenzen and Roger Wattenhofer

Page 2: MIS on Trees Christoph Lenzen and Roger Wattenhofer

Christoph Lenzen, PODC 2011Christoph Lenzen, PODC 2011

What is a Maximal Independet Set (MIS)?

• inaugmentable set of non-adjacent nodes• well-known symmetry breaking structure• many algorithms build on a MIS

Page 3: MIS on Trees Christoph Lenzen and Roger Wattenhofer

Christoph Lenzen, PODC 2011Christoph Lenzen, PODC 2011

What is a Tree?

Let’s assume we all know...

Page 4: MIS on Trees Christoph Lenzen and Roger Wattenhofer

Christoph Lenzen, PODC 2011Christoph Lenzen, PODC 2011

Talk Outline

good talkconvincing motivationimpressive resultssketch key ideascoherent conclusions

my talkWell, let’s skip that...We do it in O((ln n ln ln n)1/2) rounds!give detailsmake up for the bad talk

Page 5: MIS on Trees Christoph Lenzen and Roger Wattenhofer

Christoph Lenzen, PODC 2011Christoph Lenzen, PODC 2011

• in each phase:– draw uniformly random “ID”– if own ID is larger than all neighbors’ IDs ) join & terminate– if neighbor joined independent set ) do not join & terminate

• removes const. fraction of edges with const. probability) running time O(log n) w.h.p.

An Algorithm for General Graphs (Luby, STOC’85)

12

2

3

5

16

42

Page 6: MIS on Trees Christoph Lenzen and Roger Wattenhofer

Christoph Lenzen, PODC 2011Christoph Lenzen, PODC 2011

...and on Trees?

• same analysis gives O(log n)• ...but let‘s have a closer look:

• show that either this event is unlikelyor subtree of v contains >n nodes

survived until phase rwith degree ¢ > e(ln n ln ln n)1/2

...

...

v

Page 7: MIS on Trees Christoph Lenzen and Roger Wattenhofer

Christoph Lenzen, PODC 2011Christoph Lenzen, PODC 2011

...and on Trees?

• same analysis gives O(log n)• ...but let‘s have a closer look:

) v removed with probability¸ 1-(1-2ln ¢/¢)¢/2 ¼ 1-e-ln ¢ = 1-1/¢

survived until phase rwith degree ¢ > e(ln n ln ln n)1/2

children that surviveduntil phase r

Case 1¸ ¢/2 manywith degree · ¢/(2ln ¢)

v

Page 8: MIS on Trees Christoph Lenzen and Roger Wattenhofer

Christoph Lenzen, PODC 2011Christoph Lenzen, PODC 2011

...and on Trees?

• same analysis gives O(log n)• ...but let‘s have a closer look:

) each of them removed in phase r-1 with prob. ¸ 1-2ln ¢/¢or has ¢/(4ln ¢) high-degree children in phase r-1

survived until phase rwith degree ¢ > e(ln n ln ln n)1/2

children that surviveduntil phase r

Case 2¸ ¢/2 manywith degree ¸ ¢/(2ln ¢)

also true inphase r-1

v

Page 9: MIS on Trees Christoph Lenzen and Roger Wattenhofer

Christoph Lenzen, PODC 2011Christoph Lenzen, PODC 2011

...and on Trees?

• same analysis gives O(log n)• ...but let‘s have a closer look:

• recursion, r ¸ (ln n)1/2, and a small miracle...) v is removed in phase r with probability ¸ 1-O(1/¢)

survived until phase rwith degree ¢ > e(ln n ln ln n)1/2

children that surviveduntil phase r

...

...

v

Page 10: MIS on Trees Christoph Lenzen and Roger Wattenhofer

Christoph Lenzen, PODC 2011Christoph Lenzen, PODC 2011

Getting a Fast Uniform Algorithm

• (very) roughly speaking, we argue as follows:– degrees · e(ln n ln ln n)1/2 after O((ln n)1/2) rounds– degrees fall exponentially till O((ln n)1/2)– coloring techniques + eleminating leaves deal with small

degrees– guess (ln n ln ln n)1/2 and loop, increasing guess exponentially

) termination within O((ln n ln ln n)1/2) rounds w.h.p.

probablyO((ln n)1/2)

Page 11: MIS on Trees Christoph Lenzen and Roger Wattenhofer

Christoph Lenzen, PODC 2011Christoph Lenzen, PODC 2011

Trees - Why Should we Care?

• previous sublogarithmic MIS algorithms require small independent sets in considered neighborhood:– Cole-Vishkin type algorithms (£(log* n), directed trees, rings,

UDG‘s, etc.)– forest decomposition (£(log n/log log n), bounded arboricity)– “general coloring”-based algorithms (£(¢), small degrees)

• our proof utilizes independence of neighborsCole and Vishkin,Inf. & Control’86

Linial, SIAM J. on Comp.‘92

Schneider and Wattenhofer, PODC’08Naor, SIAM J. on

Disc. Math.‘91

Barenboim and Elkin,Dist. Comp.‘09

e.g. Barenboim and Elkin,PODC‘10

Page 12: MIS on Trees Christoph Lenzen and Roger Wattenhofer

Christoph Lenzen, PODC 2011Christoph Lenzen, PODC 2011

Some Speculation

• bounded arboricity = “everywhere sparse” ) little dependencies

) generalization possible?

• combination with techniques relying on dependence) hope for sublogarithmic solution on general graphs?

• take home message:Don‘t give up on matching the ((ln n)1/2) lower bound!

Kuhn et al., PODC’04(recently improved)

Page 13: MIS on Trees Christoph Lenzen and Roger Wattenhofer

Christoph Lenzen, PODC 2011Christoph Lenzen, PODC 2011

Thank you!