Presentation MCIP CIAC

download Presentation MCIP CIAC

of 29

Transcript of Presentation MCIP CIAC

  • 8/7/2019 Presentation MCIP CIAC

    1/29

    On the Minimum CommonInteger Partition Problem

    Author: Xin Chen, Lan Liu,

    Zheng Liu, Tao Jiang

    Presenter: Lan Liu

  • 8/7/2019 Presentation MCIP CIAC

    2/29

    Outline Introduction

    Problem definitions

    Biological applications

    Approximation of 2-MCIP

    Approximation of k-MCIP

    Conclusion and future work

  • 8/7/2019 Presentation MCIP CIAC

    3/29

    Problem Definitions P(n): given an integern, a partition is a set of

    integers, say {n1,n2,, nr}, s.t.i=1rni=n.

    Example: given n=4, {2,2} is a P(4);given n=3, {3} is a P(3).

    Observation: S= IP(S)

    Example: given S= {3, 3, 4}, {2,2,3,3} is anIP({3,3,4}).

    IP(S): given a multiset S= {x1, 0, xm}, an integerpartition is a disjoint union

  • 8/7/2019 Presentation MCIP CIAC

    4/29

  • 8/7/2019 Presentation MCIP CIAC

    5/29

    MCIP is NP-hard Subset sum P MCIP

    Subset sum problemGiven a set of integer x1, x2,, xn, s.t. X=ixi, askif

    there is a subset with the sum X/2.

    Reduction to MCIP problem

    - Let S={X/2, X/2}, T={x1, x2,, xn}, find MCIP(S,T).- If {x1, x2,, xn} is a MCIP(S,T), the answer is yes to

    Subset sum problem; otherwise, the answer is no.

  • 8/7/2019 Presentation MCIP CIAC

    6/29

    Biological Applications(1) The distance between

    two strings

    a b c d e f g h i j k hh i j k h e f g a b c d

    Genetic distance between

    two genomes

    a b c d e f g h i j k h

    h i j k h e f g a b c d

    Minimum Common

    Substring Partition

  • 8/7/2019 Presentation MCIP CIAC

    7/29

    Biological Applications(2) MCIP is a special case of Minimum Common

    Substring Partition(MCSP)

    MCIP(S',T')

    S'= {x1, x2, 0, xm}

    T'= {y1, y2, 0, yn}

    aa...a |- aa...a |- aa...ax

    1 x2 xn

    aa...a -| aa...a -| aa...ay

    1 y2 y

    MCSP(S,T)

    S=

    T=

  • 8/7/2019 Presentation MCIP CIAC

    8/29

    Outline

    Introduction

    Approximation of 2-MCIP Positive results

    Negative results

    Approximation of k-MCIP Conclusion and future work

  • 8/7/2019 Presentation MCIP CIAC

    9/29

    Some basic facts |MCIP(S1,S2,,Sk)|

    max(|S1|,|S2|,,|Sk|)

    |MCIP(S,T)| m+n-1.

    |S|=m,|T|=n

  • 8/7/2019 Presentation MCIP CIAC

    10/29

    Algorithm Analysis

    |MCIP(S,T)| m+n-1

    |MCIP(S,T)| max(m,n)

    Approximation ratio is 2

    An example: S= {3, 3, 4},T={2,2,6}

    S T CIPRound

    {1,3,4} {2,6} {2}1{3,4} {1,6} {2,1}2{2,4} {6} {2,1,1}3

    {4} {4} {2,1,1,2}4

    ; ; {2,1,1,2,4}5

    {3,3,4} {2,2,6} ;0

  • 8/7/2019 Presentation MCIP CIAC

    11/29

    Definitions for MRSP(1) Related multisets: ifS=Tand S,T{;, Sand T

    are a pair of related multisets.

    Example:{3

    3

    4

    5

    10}{2 2 687}

    {334510}

    {22687}

    Basic related multisets: if there are no S' Sand T' T, s.t. S'and T'are related.

    Example:

  • 8/7/2019 Presentation MCIP CIAC

    12/29

    Definitions for MRSP(2) Maximum Related Multiset Partition problem(MRSP)

    Given S and T, partition them into related submultisets

    with the maximum cardinality.

    (2){334 10}

    {22687}

    (1){334 10}

    {22687}

    (3){334 10}

    {22687}

    Observation: IfS, Tare

    a pair of basic related

    multisets, |MRSP|=1.

  • 8/7/2019 Presentation MCIP CIAC

    13/29

    MRSP $ 2-MCIP CIP! RSP

    S: {4 3 35 10}

    T: {2 2 687}

    CIP: {2 2 3 353 7}

    For each component,

    #edges #vertices 1

    Each component is

    related.

    S: {4 3 35 10}

    T: {22 687}

    |CIP| m+n-|RSP|

    |MCIP| m+n-|RSP|

    m+n-|MRSP|

    S: {4 33510}

    T: {22 687}

  • 8/7/2019 Presentation MCIP CIAC

    14/29

  • 8/7/2019 Presentation MCIP CIAC

    15/29

    MRSP $ 2-MCIP |MRSP| = m+n|MCIP|

    IfS, Tare a pair of basic related multisets,

    |MCIP|= m+n-1, because |MRSP|=1. When m+n 5, |MCIP| =m+n-1 4/5(m+n).

    A new way to solve MCIP

    Step1. find MRSP;

    Step2. for each basic related submultiset, run

    Greedy_CIP(S', T').

  • 8/7/2019 Presentation MCIP CIAC

    16/29

    Approximate 2-MCIP Algorithm intuition:

    Step 1. find related submulitsets

    Step 2. set packing

    Step 3. Greedy-CIP

    mimic MRSP

  • 8/7/2019 Presentation MCIP CIAC

    17/29

    Set Packing Problem(1) Set Packing

    Given a set of subsets S, find the largest number of

    mutually disjoint subsets from S?

  • 8/7/2019 Presentation MCIP CIAC

    18/29

    Set Packing Problem(2) Bad news

    - It is NP-hard to find related submultisets oflarge size.

    - Set packing itself is NP-hard.

    Good news

    We can find the small related submultisets andapproximate set packing efficiently.

  • 8/7/2019 Presentation MCIP CIAC

    19/29

    Approximate 2-MCIP Main idea: use different strategies for the

    submultisets with different sizes.

    The approximation ratio is 5/4.

    If there are no basic related submultisets with sizesmaller than 5, 4/5 (m+n) |MCIP| m+n-1.

    Str t i sR v _c _i t r

    r xi t _s t_ cki

    Submultis t siz

    ,r mor r y_ I

  • 8/7/2019 Presentation MCIP CIAC

    20/29

    Outline

    Introduction

    Approximation of 2-MCIP Positive results

    Negative results

    Approximation ofk-

    MCIP

    Conclusion and future work

  • 8/7/2019 Presentation MCIP CIAC

    21/29

    General frameworkx f(x )

    g(x ,y) y

    IP

    IP

    SOLP

    (x ) SOLP

    (f(x ))

    LinearReduction L

    OPTP2(f(x)) E OPTP1(x)

    |OPTP1(x)- g(x,y)| F|OPTP2(f(x))-y|

    IfP1 cannot be approximated

    within some constant ratio c,

    P2 cannot be approximated by

    some constant ratio c'.

  • 8/7/2019 Presentation MCIP CIAC

    22/29

    Maximum 3DM-3 Problem Definition

    Given a set D XYZ, where X, Yand Z are disjoint sets,

    and each element occurs in at most three triples, find amatching with the maximum cardinality.

    Known fact

    Maximum 3DM-3 cannotbe approximated within some

    constant ratio. [Kann91]

    X:

    Y:

    Z:

    X:

    Y:

    Z:

  • 8/7/2019 Presentation MCIP CIAC

    23/29

    L-reduction(1) f: S={4i|i2X[Y[Z }

    T={4i1+4i2+4i3|(i1,i2,i3)2D}

    D

    X Y ZS :

    T:

    OPTMCIP

    70*OPT3DM

  • 8/7/2019 Presentation MCIP CIAC

    24/29

    g: - CIP ! RSP|OPTRSPSOLRSP| |OPTMCIPSOLMCIP|

    - RSP ! 3DM

    X

    S1:

    Y Z

    T1:

    10 1... ...

    D

    *

    1 1

    * *

    di

    i1 i3i2

    OPT3DMOPTRSP

    Each related submultiset

    includes at least one triple

    L-reduction(2)

    |OPT3DMSOL3DM| |OPTRSPSOLRSP|

  • 8/7/2019 Presentation MCIP CIAC

    25/29

  • 8/7/2019 Presentation MCIP CIAC

    26/29

    Outline

    Introduction

    Approximation of 2-MCIP

    Approximation of k-MCIP

    Conclusion and future work

  • 8/7/2019 Presentation MCIP CIAC

    27/29

    Approximate k-MCIP Run Greedy_CIP(S,T) sequentially on S1,S2, ,

    Sk.

    |MCIP(S1,S2,,Sk)| |S1|+|S2|++|Sk|

    |MCIP(S1,S2,,Sk)| max(|S1|,|S2|,,|Sk|)

    Approximation ratio is k

    We can get a {3k(k-1)}/(3k-2)- approximation

    by removing the common elements.

  • 8/7/2019 Presentation MCIP CIAC

    28/29

    Outline

    Introduction

    Approximation of 2-MCIP Approximation of k-MCIP

    Conclusion and future work

    Upper bound5/4

    {3k(k-1)}/(3k-2)2-MCIP

    k-MCIP (k>2)

    Lower boundAPX-hardAPX-hardAPX-hard

  • 8/7/2019 Presentation MCIP CIAC

    29/29

    Thanks for your time and

    attention!