On the Complexity of K-Dimensional-Matching Elad Hazan, Muli Safra & Oded Schwartz
Maximal Matching in Bipartite Graphs
Easy problem: in P Maximal Matching in Bipartite Graphs
3-Dimensional Matching (3-DM)
NP-hard [Karp72] 3-Dimensional Matching (3-DM) Matching in a bounded hyper-graph Bounded Set Packing
3-DM: Bounded Set-Packing Maximal Matching in a Hyper-Graph which is 3-uniform & 3-strongly-colorable Set-Packing: [BH92] [Hås99] Bounded variant: App. : [HS89] Inapp. : [CC03]
K
K
k-DM: Bounded Set-Packing Maximal Matching in a Hyper-Graph which is k-uniform & k-strongly-colorable Set-Packing: [BH92] [Hås99] Bounded variant: App. : [HS89] Inapp. : [Tre01] Without this this is k-SP
Unless P=NP, k-DM cannot be approximated to within Main Theorem: Corollary: The same holds for k-Set-Packing and Independent set in k+1-claw-free graphs Some inapproximability factors for small k-values are also obtained
Gap-Problems and Inapproximability Maximization problem A Gap-A-[s no, s yes ]
Gap-Problems and Inapproximability Maximization problem A Gap-A-[s no, s yes ] is NP-hard. Approximating A better than s yes /s no is NP-hard.
Gap-Problems and Inapproximability Gap-k-DM-[ ] is NP-hard. k-DM is NP-hard to approximate to within
L-q: Input: A set of linear equations mod q Objective: Find an assignment satisfying maximal number of equations App. ratio: 1/q Inapp. factor: 1/q+ [Hås97] x 1 + x 2 + x 3 = a 1 mod q x 7 + x 4 + x 2 = a 2 mod q … x 8 + x 2 + x 9 = a n mod q
Thm [Hås97]: Gap-L-q-[1/q+ , 1- ] is NP-hard. Even if each variable x occurs a constant number of times, c x = c x ( ) x 1 + x 2 + x 3 = a 1 mod q x 7 + x 4 + x 2 = a 2 mod q … x 8 + x 2 + x 9 = a n mod q
Gap-L-q ≤ p Gap-k-SP x 1 + x 2 + x 3 = a 1 mod q x 7 + x 4 + x 2 = a 2 mod q … x 8 + x 2 + x 9 = a n mod q Can be extended to k-DM
Gap-L-q ≤ p Gap-k-SP H = (V,E) We describe hyper edges, then which vertices they include. x 1 + x 2 + x 3 = a 1 mod q x 7 + x 4 + x 2 = a 2 mod q … x 8 + x 2 + x 9 = a n mod q 1 st trial:
Gap-L-q ≤ p Gap-k-SP A hyper-edge for each equation and a satisfying assignment to it (q 2 such assignments). x 1 + x 2 + x 3 = a 1 mod q x 7 + x 4 + x 2 = a 2 mod q … x 8 + x 2 + x 9 = a n mod q 1 : x 1 + x 2 + x 3 = 0 mod 3 A( 1 )=(0,1,2) 2 : x 7 + x 4 + x 2 = 1 mod 3 A( 2 )=(1,0,0)
1 st trial: Gap-L-q ≤ p Gap-k-SP A hyper-edge for each equation and a satisfying assignment to it A common vertex for each two contradicting edges x 1 + x 2 + x 3 = a 1 mod q x 7 + x 4 + x 2 = a 2 mod q … x 8 + x 2 + x 9 = a n mod q 1 : x 1 + x 2 + x 3 = 0 mod 3 A( 1 )=(0,1,2) 2 : x 7 + x 4 + x 2 = 1 mod 3 A( 2 )=(1,0,0) x 2 :(1,0)
1 st trial: Gap-L-q ≤ p Gap-k-SP Maximal matching Consistent assignment x 1 + x 2 + x 3 = a 1 mod q x 7 + x 4 + x 2 = a 2 mod q … x 8 + x 2 + x 9 = a n mod q 1 : x 1 + x 2 + x 3 = 0 mod 3 A( 1 )=(0,1,2) 2 : x 7 + x 4 + x 2 = 1 mod 3 A( 2 )=(1,0,0) x 2 :(1,0)
1 st trial: Gap-L-q ≤ p Gap-k-SP Maximal matching Consistent assignment Gap-L-q-[1/q+ ,1- ] < p Gap-k-SP-[1/q+ ,1- ] What is k ? x 1 + x 2 + x 3 = a 1 mod q x 7 + x 4 + x 2 = a 2 mod q … x 8 + x 2 + x 9 = a n mod q k is large ! k (c x1 +c x2 +c x3 ) q(q-1)
Gap-L-q ≤ p Gap-k-SP Saving a factor of q: Reuse vertices k Still depends on c x1 +c x2 +c x3 which depends on x 1 + x 2 + x 3 = a 1 mod q x 7 + x 4 + x 2 = a 2 mod q … x 8 + x 2 + x 9 = a n mod q x 2 =1 x 2 =2 x 2 =0
2 nd trial: Gap-L-q ≤ p Gap-k-SP Allow pluralism: A (few) contradicting edges may reside in a matching Common vertices for only some subsets of contradicting edges - using a connection scheme. x 1 + x 2 + x 3 = a 1 mod q x 7 + x 4 + x 2 = a 2 mod q … x 8 + x 2 + x 9 = a n mod q
Which contradicting edges to connect ? A Connection Scheme for x cxcx q Fewer vertices: Consistency achieved using disperser-Like Properties
Def:[HSS03] -Hyper-Disperser H=(V,E) V=V 1 V 2 … V q E V 1 × V 2 × … × V q U independent set (of the strong sense) i, |U\V i | < |V| If U is large it is concentrated ! This generalizes standard dispersers
Lemma [HSS03]: Existence of -Hyper-Disperser q>1,c>1 1/q 2 -Hyper-Disperser which is also q uniform, q strongly-colorable d regular, d strongly-edge-colorable for d= (q log q) Proof … Optimality …
Def:[HSS03] -Hyper-Edge-Disperser H=(V,E) E=E 1 E 2 … E q M matching i, |M\E i | < |E| If M is large it is concentrated !
Lemma [HSS03]: Existence of -Hyper-Edge-Disperser q>1,c>1 1/q 2 -Hyper-Edge-Disperser which is also q regular, q strongly-edge-colorable d uniform, d strongly-colorable for d= (q log q) Jump …
Constructing the k-SP instance H =(V,E) x - a copy of (c=c x ). V the vertices of all
E for each equation and a satisfying assignment to it – the union of three hyper-edges : x 1 + x 2 + x 3 = 4 A( )=(0,1,3) X1X1 X3X3 X2X2 e ,(0,1,2) Constructing the k-SP instance H =(V,E) H is 3d uniform 3d= (q log q)
Completeness: If A satisfying 1- of then M covering 1- of V (hence of size |V|/k) Proof: Take all edges corresponding to the satisfying assignment. ڤ
Soundness: If A satisfies at most 1/q + of then M covers at most 4/q 2 + of V
Soundness-Proof: M maj Edges of M that agree with A M min M \ M maj (Håstad) A most popular values of each
Soundness-Proof: Every edge of M min is a minority in at least one
Soundness-Proof:
Gap-L-q-[1/q+ ,1- ] ≤ p Gap-k-SP- [O(1/q),1- ] What is k ? Gap-k-SP-[ ] is NP-hard. Unless P=NP, k-SP cannot be approximated to within k=3d= (q log q)
Conclusion Unless P=NP, k-SP cannot be approximated to within This can be extended for k-DM. 4-DM, 5-DM and 6-DM cannot be approximated to within respectively. Deterministic reduction
Open Problems Low-Degree: 3-DM,4-DM… TSP Steiner-Tree Sorting By Reversals
Open Problems Separating k-IS from k-DM ? k-DMk-IS App. ratio Innap. factor [Vis96] [Tre01][HSS03] [HS89]
THE END
Optimality of Hyper-Disperser: 1/q2-Hyper-Disperser Regularity: d= (q log q) Restrict hyper disperser to V 1,V 2. A bipartite -Disperser is of degree (1/ log 1/ ) and 1/q. Definition …
Existence of Hyper-Disperser Proof: random construction. Random permutations: j i R S c j {2,…,q}, i [d] e[i,j] = { v[1,j], v[2, 2 i (j)], …, v[q, k i (j)] } E = {e[i,j] | j {2,…,q}, i [d] } Definition …
Proof – cont. Candidates: ‘bad’ (minimal) sets: U = { U | U V, |U| = 2c/q, |U V 1 |=c/q}
Proof – cont.
Extending it to k-DM Gap-k-SP-[O(log k / k), 1- ] is NP-hard.
Use a for each location of a variable. Gap-k-DM-[O(log k / k), 1- ] is NP-hard.
From Asymptotic to Low Degree – How to make k as small as possible ? Minimize d ( = 3) – by minimizing q ( = 2) (a bipartite disperser) Avoid union of edges
E equation and a satisfying assignment to it –three hyper-edges : x 1 + x 2 + x 3 = 0 A( )=(0,1,1) X1X1 X3X3 X2X2 e ,(0,1,2),x1 e ,(0,1,2),x2 e ,(0,1,2),x3 From Asymptotic to Low Degree – How to make k as small as possible ?