week 9, euclidean tsp - vuarora’s ptas de nition a polynomial time approximation scheme (ptas)...

32
Week 9, Euclidean TSP Randomized algorithms

Upload: others

Post on 18-Mar-2021

2 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Week 9, Euclidean TSPRandomized algorithms

Page 2: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Today

1. Few’s algorithm (1955)

2. Karp’s algorithm (1977)

3. Arora’s algorithm (1998)

Page 3: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Arora’s PTAS

DefinitionA polynomial time approximation scheme (PTAS) gives for anyfixed ε > 0 a polynomial time (1 + ε)-approximation algorithm.

Theorem (Arora, 1998)

There is a PTAS for the TSP in the Euclidean plane.

Sanjeev Arora: Polynomial Time Approximation Schemes for EuclideanTraveling Salesman and other Geometric Problems. J. ACM 45 (1998)- Cited 1097 times (May 2019, Google)

Page 4: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

1. Few’s algorithm (1955)

Page 5: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Few’s algorithm

Page 6: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Few’s algorithm

TheoremFor any set of n point in a 1× 1 square, there exists a TSP tour oflength O(

√n).

Page 7: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

2. Karp’s algorithm (1977)

Page 8: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Karp’s algorithm

Page 9: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Karp’s algorithm

Page 10: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Karp’s algorithm

TSP solvable in O(n32n) time.Choose s = log2 n.

TheoremKarp’s algorithm finds a tour of length O(

√(n/ log2 n)) + OPT .

Page 11: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Random points

Place n points uniformly at random in a 1× 1 square.

TheoremThe expected length is Θ(

√n)

Corollary

For n random points:

1 6ZKARP

Opt6 1 + o(1).

(i.e., Karp is optimal in the limit for random points.)

Page 12: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

3. Arora’s algorithm (1998)

Page 13: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Euclidean TSP

Theorem (Arora, 1998)

There is a PTAS for the TSP in the Euclidean plane.

Main ingredients

I Roundig the instance.

I Restricting the solution space.

I Dynamic programming.

I Randomization.

Page 14: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Euclidean TSP

The algorithm

1. Take a random box B that covers all points.

2. Move points to grid points.

3. Build quad tree.

4. Define portals.

5. Define the subproblems for the DP.

6. Fill the DP table.

Page 15: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Step 1, Random box B

Let k = dlog2(n/ε)e. Take a, b ∈ {0, 1, . . . , 2k−1 − 1} at random.

Page 16: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Step 2, Rounding

Move each point to the middle of its 1× 1 cell.

Page 17: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Step 3, Quad trees

Iteratively divide the square in 4 squares until smallest are 1× 1.

Page 18: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Step 4, Portals

Each gridline gets a number of regularly distributed points.

Page 19: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Step 5, Subproblems

A subproblems is given by (X ,Y ,Z ):

X : A square of the quadtree.

Y : An even subset of portals of X .

Z : A pairing of the portals Y .

Store in the D.P. :F (X ,Y ,Z ) = minimum length of set of paths inside X that crossthe boundary at Y and obey the pairing Z .

Page 20: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Step 6, Fill the DP table.

A value F (X ,Y ,Z ) can be computed from all possible valuesF (X1,Y1,Z1), F (X2,Y2,Z2), F (X3,Y3,Z3), and F (X4,Y4,Z4),where X1,X2,X3, and X4 are the childeren of X .

Page 21: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Analysis

Need to show that the

1 approximation ratio is at most 1 + ε, and

2 running time is polynomial.

Page 22: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Analysis

1 approximation ratio.

Page 23: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Analysis, approx. ratio

Rounding. Note that

I Opt > L = 2k > n/ε (k = dlog2(n/ε)e.)I Difference per point 6

√2

Rounding error: 6√

2n 6√

2εL 6√

2εOpt.

Let Opt′ be optimal value for rounded instance

Opt′ 6 (1 +√

2ε)Opt

Page 24: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Analysis, approx. ratio

Restriction to portals.Let δi be interportal distance of a level i line.

δi =L

m2i=

2k

m2i.

Detour per crossing: 6 δi .

May be too large if random (a, b)-shift was not used!!!

Page 25: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Analysis, approx. ratio

Restriction to portals.

Let l be arbitrary grid line.Pr(l is of level 1) = 21−k .Pr(l is of level 2) = 21−k .Pr(l is of level 3) = 22−k ....Pr(l is of level i) = 2i−1−k , for i > 2.⇒Pr(l is of level i) 6 2i−k , for i > 1.

Page 26: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Analysis, approx. ratio

Restriction to portals.

Expected length of detour for one crossing with gridline l :

6k∑

i=1

Pr(l is of level i) · δi

6k∑

i=1

2i−k · 2k

m2i=

k

m=

k

dk/εe6 ε.

Choose m = dk/εe = ddlog2(n/ε)e/εe = O( log nε ).

Page 27: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Analysis, approx. ratio

Claim: At most√

2Opt′ crossings with gridlines in total.

Total detour for portals ε√

2Opt′.

Let Opt′′ be optimal value for a portal respecting tour for roundedinstance

E[Opt′′] 6 (1 + ε√

2)Opt ′

Page 28: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Analysis, approx. ratio

Summarizing:

Opt is optimal length original instance and,Opt′ is optimal for rounded instance and,Opt′′ is optimal portal respecting tour for rounded instance.

E[Opt′′] 6 (1 + ε√

2)Opt ′

6 (1 + ε√

2)(1 +√

2ε)Opt

= (1 + O(ε))Opt.

Page 29: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Analysis

2 running time.

Page 30: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Analysis, running time

� Number of subproblems (X ,Y ,Z ):I X : Size of quadtree: 4k leaves. ⇒ O(n2/ε2) squares X .I Y : nO(1/ε) ways to cross portals at X .I Z : nO(1/ε) ways to pair crossings of Y .

⇒ Number of subproblems (X ,Y ,Z ) = nO(1/ε)

� Computing one value F (X ,Y ,Z ) takes nO(1/ε) time.

Total time and space is nO(1/ε).

Page 31: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Extensions

This technique has been used to get a PTAS for the Euclideanversion of

I Minimum Steiner tree.

I Mimimum k median.

I k-TSP

I k-MST

I mincost k-connected subgraph.

I Traveling Repairman Problem

I many others ....

Page 32: Week 9, Euclidean TSP - VUArora’s PTAS De nition A polynomial time approximation scheme (PTAS) gives for any xed >0 a polynomial time (1 + )-approximation algorithm. Theorem (Arora,

Some open problems

Only Quasi-PTAS known: (nO(log n) time)

I Capacitated vehicel routing. (QPTAS by A. Das and C.Mathieu, 2010)

I Mimimum weight triangulation (QPTAS by Remy and Steger,2006)

I Prize collecting TSP (QPTAS by Arora)

No scheme known:

I Mimimum weight Steiner triangulation.