minimum back-walk-free latency problem yaw-ling lin dept computer sci. & info. management,...

26
Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan.

Post on 15-Jan-2016

214 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Minimum Back-Walk-Free Latency Problem

Yaw-Ling Lin

Dept Computer Sci. & Info. Management,Providence University, Taichung, Taiwan.

Page 2: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 2

Minimum Latency Problem (MLP)

• Starts from s, sending goods to all other nodes.

• Traveling Salesperson Problem (TSP): Server oriented

• MLP: Client oriented

• MLP is also known as repairman problem or traveling repairman problem (TRP)

s

Page 3: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 3

MLP: Formal Definition

Page 4: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 4

MLP vs. TSP• TSP: minimizes the salesman’s total time. Server oriented, egoistic.

– No contstant approximation algorithm for general case.– Christofides (1976): 3/2-approximation ratio for metric case; Arora (1992):

metric TSP does not have PTAS unless P=NP.– Arora (1998 JACM): PTAS on Euclidean case.

• MLP: minimizes the customers’ total time. Clients oriented, altruistic.– Alias: deliveryman problem, traveling repairman problem (TRP).– Afrati (1986): MAX-SNP-hard for metric case.– Goeman (1996): 10.78-approximation ratio for metric case (with Garg, 19

96FOCS, technique); 3.59-approximation ratio for trees.– Arora (1999 STOC): quasi-polynomial ( O(nO(log n) ) approximation scheme

for trees and Euclidean space. – Sitters (2002, IPCO): MLP on trees is NP-complete; not known for caterpi

llars.

Page 5: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 5

MBLP: Back-Walk Free

Page 6: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 6

An Example

Page 7: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 7

Our Results

• MBLP, given a starting point of G– Trees : O(n log n ) time– k-path : O(n log k) ; path is O(n) time– DAG : NP-Hard (Reduce from 3-SAT)

Page 8: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 8

Properties of MBLP on Trees

Page 9: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 9

Properties (contd’)

Page 10: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 10

Properties (3)

Page 11: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 11

Algo MBLP-Tree: Example

5

3 10

8 11 2

Select 11

5

14/2 10

8 2

Select 10

15/2

14/2

8 2

Select 8

15/2

22/3 2

Select and output 15/222/3 2

Select and output 22/3

2Select and output 2

Result : 5,10,3,11,8,2

Page 12: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 12

Algorithm MBLP-Tree

Page 13: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 13

Analysis of MBLP-Tree

Page 14: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 14

Properties of MBLP on k-Path

<4,2,3,8> is right-skew; < 5, 3, 4, 1, 2, 6 > is not.<5> <3,4> <1,2,6> is decreasing right-skew partitioned.

Page 15: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 15

Properties of k-Path (contd’)

Page 16: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 16

Path-Partition: Example

Page 17: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 17

Algorithm Path-Partition

Page 18: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 18

Algorithm k-Path

Page 19: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 19

Analysis of k-Path

Page 20: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 20

MBLP on DAG is NP-Complete

Page 21: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 21

NP-Completeness: Construction

-- Reduction from 3SAT: n literals:

S

x1 x1

x2 x2

x3 x3

x4 x4

… n literals

1 literal

Page 22: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 22

NP-Completeness: Construction

-- Reduction from 3SAT: k clauses:

0 0 0

0 0 0

… k clauses

1 caluse

Page 23: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 23

NP-Completeness: illustration

Page 24: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 24

NP-Complete Proof

Page 25: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 25

Conclusion

• MBLP is easier than MLP, at least on trees.• MBLP remains hard even on dag.• The idea of atomic subtours helps in finding effici

ent algorithms of MBLP on trees.• The idea of atomic sequence becomes right-skew

partition, implying the linear time algorithm on paths.

Page 26: Minimum Back-Walk-Free Latency Problem Yaw-Ling Lin Dept Computer Sci. & Info. Management, Providence University, Taichung, Taiwan

Yaw-Ling Lin, Providence, Taiwan 26

Future Research

• MLP on caterpillars.

• MBLP: finding the good starting points on paths, trees.

• MBLP: multiple servers on trees, paths.