• No results found

shayan.pptx

N/A
N/A
Protected

Academic year: 2020

Share "shayan.pptx"

Copied!
31
0
0

Loading.... (view fulltext now)

Full text

(1)

Effective-Resistance-Reducing Flows,

Spectrally Thin Trees, and ATSP

Nima Anari UC Berkeley

(2)

Asymmetric TSP (ATSP)

Given a list of cities and their pairwise β€œdistances”, satisfying the triangle inequality,

Find the shortest tour that

visits all cities exactly once.

(3)

Linear Programming Relaxation

[Held-Karp’72]

π‘šπ‘–π‘›

βˆ‘

𝑖

,

𝑗

𝑐

(

𝑖

,

𝑗

)

π‘₯

𝑖

,

𝑗

ΒΏ

𝑠

.

𝑑

.

βˆ‘

𝑗

π‘₯

𝑖

,

𝑗

=

βˆ‘

𝑗

π‘₯

𝑗

,

𝑖

βˆ€π‘–βˆˆπ‘‰

βˆ‘

π‘–βˆˆπ‘†

,

π‘—βˆ‰π‘†

π‘₯

𝑖

,

𝑗

β‰₯

1

ΒΏ

0

≀

π‘₯

𝑖

,

𝑗

ΒΏ

ΒΏ

(4)

Previous Works

Approximation Algorithms

β€’ log(n) [Frieze-Galbiati-Maffioli’82]

β€’ .999 log(n) [BlΓ€ser’02]

β€’ 0.842 log(n) [Kaplan-Lewenstein-Shafrir-Sviridenko’05]

β€’ 0.666 log(n) [Feige-Singh’07]

β€’ O(logn/loglogn) [Asadpour-Goemans-Madry-O-Saberi’09]

β€’ O(1) (planar/bd genus) [O-Saberi’10,Erickson-Sidiropoulos’13]

Integrality Gap

(5)

Main Result

For any cost function, the integrality gap of the LP relaxation is polyloglog(n).

(6)

Plan of the Talk

ATSP

ATSP Thin Spanning

(7)

Thin Spanning Trees

Def: Given a k-edge-connected graph . A spanning tree is -thin w.r.t. G if

Kn 2/n-thin tree

One-sided unweighted cut-sparsifier No lower-bound on

One-sided unweighted cut-sparsifier No lower-bound on

βˆ€ 𝑆 βŠ†π‘‰ ,

|

𝑇

(

𝑆 , 𝑆

)

|

≀ 𝛼 β‹…βˆ¨πΈ

(

𝑆 ,𝑆

)

∨¿

Exercise: Show that (k-dim cube) has O(1/k) thin tree

(8)

From Thin Trees to ATSP

[AGMOS’09]: If for any -connected graph , then the integrality gap of LP is .

Furthermore, finding the tree algorithmically gives -approximation algorithm for ATSP.

(9)

Previous Works: Randomized Rounding

Thm: Any k-connected graph G has a thin tree

Pf. Sample each edge of G, indep, w.p. .

By Karger’s cut counting argument, the sampled graph is -thin w.h.p.

(10)

Main Result

Any -edge-connected graph has an -thin tree.

Any -edge-connected graph has an -thin tree.

For any cost function, the integrality gap of the LP Relaxation is polyloglog(n).

(11)

In Pursuit of Thin Trees

(12)

Graph Laplacian

Let

For let

Laplacian Quadratic Form:

𝐿𝐺=

[

2 βˆ’ 1 βˆ’1 0

βˆ’1 2 0 βˆ’1

βˆ’1 0 2 βˆ’1 0 βˆ’ 1 βˆ’1 2

]

(13)

Spectrally Thin Spanning Trees

Def: A spanning tree is -spectrally thin w.r.t. G if

Why?

β€’ Generalizes (combinatorial) thinness.

– -spectral thinness implies -combinatorial thinness

β€’ Testable in polynomial time.

(14)

Lem: The spectral thinness of any T is at least

Pf. If T is spectrally thin, then any subgraph of T is -spectrally thin, so

is the spectral thinness of .

A Necessary Condition for Spectral Thinness

(15)

A k-con Graph with no Spectrally Thin Tree

For any spanning tree, T,

Reff (𝑒)β‰ˆ 1βˆ’ π‘˜

2

𝑛

k edges k edges

(16)

A Sufficient Condition for Spectral Thinness

[Marcus-Spielman-Srivastava’13,Harvey-Olver’14]: Any G has an -spectrally thin tree.

(17)
(18)
(19)

Main Idea

(20)

An Example

(21)

An Observation

An Application of [MSS’13]: If for any cut ,

(22)

Main Idea

Find a ``graph’’ D s.t. and

i.e.,

Find a ``graph’’ D s.t. and

i.e.,

D+G has a spectrally thin tree and

any spectrally thin tree of G+D is (comb) thin in G. D+G has a spectrally thin tree and

(23)

An Impossibility Theorem

(24)

Proof Overview

k-connected graph G for

k-connected graph G for

,

F is -connected, ,

F is -connected,

G has -comb thin tree G has -comb

thin tree

A General. of [MSS’13] Main

Tech Thm

has -spectrally thin tree

has -spectrally thin tree

Note we may have Note we may have D is not

a graph D is not

(25)

Thm: Given a set of vectors s.t.

If then there is a basis s.t.

………...…,there are disjoint bases,

Thin Basis Problem

β€–

βˆ‘

𝑒 βˆˆπ‘‡

𝑣

𝑒

𝑣

𝑒𝑇

β€–

≀

𝑂

(

πœ–

)

.

d Linearly independent set of vectors

(26)

Proof Overview

k-connected graph G for

k-connected graph G for

,

F is -connected, ,

F is -connected,

G has -comb thin tree G has -comb

thin tree

A General. of [MSS’13] Main

Tech Thm

has -spectrally thin tree

(27)

A Weaker Goal: Satisfying Degree Cuts

Thm: Given a k-connected graph, , s.t., for all v,

for

Let then by Markov Ineq,

, for all v.

(28)

A Convex Program for Optimum D

π‘šπ‘–π‘›

max

π‘£βˆˆπ‘‰

𝔼

π‘’βˆΌπ›Ώ(𝑣)

Reff

𝐷

(

𝑒

)

ΒΏ

𝑠

.

𝑑

.

𝐷

β‰Ό

𝐢

𝐿

𝐺

ΒΏ

ΒΏ

Has exp. many constraints: Has exp. many constraints:

Recall convexity of matrix inv Recall convexity of matrix inv

If write ,

then optimum is If write ,

then optimum is

(29)

Main Result

Any -edge-connected graph has an -thin tree. Any -edge-connected graph has an -thin tree.

For any cost function, the integrality gap of the LP Relaxation is polyloglog(n).

(30)

Conclusion

Main Idea:

β€’ Symmetrize L2 structure of G while preserving its L1 structure

Tools:

β€’ Interlacing polynomials/Real Stable polynomials β€’ Convex optimization

β€’ Graph partitioning

(31)

Future Works/Open Problems

β€’ Algorithmic proof of [MSS’13] and our extension.

β€’ Existence of C/k thin trees and constant factor

approximation algorithms for ATSP.

β€’ Subsequent work: Svensson designed a 27-app

References

Related documents