Julia Chuzhoy (TTI-C) Yury Makarychev (TTI-C) Aravindan Vijayaraghavan (Princeton) Yuan Zhou (CMU) Cut minimization • Min st-cut: delete the min #edges to disconnect s, t t s Duality: Maxflow(s, t) = Mincut(s, t) =2 Multicut • Given r pairs (si, ti), delete min #edges to disconnect all (si, ti) pairs t 1 •Upper bound on max multicommodity flow t3 s3 •Identifies bottlenecks in the graph s1 t2 s2 •O(log r) approximation algorithm [GVY95] Min k-route cuts • Unweighted version. Given r pairs (si, ti), delete min #edges to k-disconnect all (si, ti) pairs – i.e. for all i, (si, ti)-edge-connectivity < k • General version. Given a weighted graph and r pairs, delete min wt. of edges to k-disconnect all (si, ti) pairs t1 t3 s3 s1 t2 s2 For example, when k = 2, OPT = 1. Min k-route cuts: variants and specal cases • EC-kRC: edge connectivity version, remove min. wt. of edges so that for each i, (si, ti)-edge-connectivity < k – Unweighted case: all edge weights = 1 – k = 1: Minimum multicut • VC-kRC: vertex connectivity version, remove min. wt. of edges so that for each i, (si, ti)-vertex-connectivity < k Motivation • Multiroute generalization : a fault tolerant setting • st-k-route flow: a fractional combination of elementary k-route st-flows [Kis96, KT93, AO02] – Flow is resilient to (k-1) failures Maxflow/ Mincut multiroute generalization st-k-route flow multicommodity flow k-route multicommodity flow multicut k-route cut Motivation (cont'd) • Multiroute generalization: a fault tolerant setting • As standard multicut, k-route cut also reveals network bottleneck, and in particular measures resilience of the network multicut k-route cut Approximation algorithms • α-approximation: delete edges of wt. αOPT such that all the pairs are k-disconnected • (β,α)-bicriteria approximation: delete edges of wt. αOPT such that all the pairs are βk-disconnected Previous work • [Chekuri-Khanna'08] – O(log2n log r)-approximation for k=2 (both EC-2RC and VC-2RC) • [Barman-Chawla'10] – O(log2r)-approximation for k=2 (both EC-2RC and VC-2RC) • [Kolman-Scheideler'11] – O(log3r)-approximation for k=3 (EC-2RC) • No sub-polynomial approx. algorithm known for k > 3 Our results : algorithms for EC-kRC • Unweighted EC-kRC – O(k log1.5 r)-approximation – (1+ε, (1/ε)log1.5 r)-bicriteria approximation • General EC-kRC – O(log1.5 r)-approximation for k = 2 – (2, log2.5 r loglog r)-bicriteria approx. in nO(k) time – (log r, log3 r)-bicriteria approx. in poly(n, k) time Our results : VC-kRC • Algorithms – O(log1.5 r)-approximation for k = 2 – (2, d k log2.5 r loglog r)-bicriteria approx. in nO(k) time, where each node belongs to at most d source-sink pairs • Harndess for VC-kRC – NP-Hard to approximate VC-kRC within Ω(kε) for some specific ε > 0 • Hardness for st-VC-kRC – Superconstant hardness assuming random k-AND hypothesis of [Feige'02] – Ω(ρ0.5) hardness assuming ρ-inapproximability of Densest k-Subgraph A comparison : EC-kRC Previous results Our results k=2 O(log2 r) [BC10] O(log1.5 r) k=3 O(log3 r) [KS11] arbitrary k, unweighted arbitrary k, general O(k log1.5 r) (1+ε, (1/ε)log1.5 r) (2, log2.5 r loglog r) in time nO(k) (log r, log3 r) in poly(n, k) time A comparison : VC-kRC k=2 arbitrary k Previous results Our results O(log2 r) [BC10] O(log1.5 r) multicut hardness: 2.5r loglog r) (2, dklog APX-hard [DJP+94] alg in time nO(k) superconstant assuming UGC ε)-hardness Ω(k [KV05, CKK+06] The rest of this talk... • O(k log1.5 r)-approximation algorithm for unweighted ECkRC • (2, log2.5 r loglog r)-bicriteria approx. algorithm for general EC-kRC (sketch) The difficulty for large k (> 2) • Simple recursion (used in [BC10]) for k = 2 – Find a balanced cut (by region growing) – Remove all the cut edges but the most expensive one – recurse into both sides s 1 • Key observation. the red edge cannot provide extra connectivity for s1, t1 t1 graph G The difficulty for large k (> 2) • Simple recursion (used in [BC10]) for k = 2 – Find a balanced cut (by region growing) – Remove all the cut edges but the most expensive one – recurse into both sides s graph G 1 • Key observation. the red edge cannot provide extra connectivity for s1, t1 t1 a bad example for k = 3 • No longer true for k = 3 (or more) Algorithms for k > 2 • [Kolman-Scheideler'11] O(log3r)-approximation for k=3, by multi-level region growing (based on the same LP used in [BC10]) • Our method – Idea 1. Relate k-route cut to the value of sparest cut – Idea 2. Solve the problem iteratively rather than recursively O(k log1.5 r)-approximation algorithm for unweighted EC-kRC Cut sparsity, and unweighted EC-kRC • Let d(v) = #source-sink pairs that v participates in d (v ) d(S) = vS • Define uniform sparsity to be ( S ) edges( S , S ) min{d ( S ), d ( S )} , (G ) min ( S ) S • Theorem.[ARV04] O(log0.5 r)-approx. for Φ(G). • Lemma. (G ) k OPT r Algorithm for unweighted EC-kRC • Step 0. Assume source-sink pairs are not k-disconnected • Step 1. Use the algorithm in [ARV04] to find an approximate sparse cut (S , S ) • Step 2. Delete all the edges across the cut • Step 3. Recurse into the subinstances defined by each side of the cut k OPT • Lemma. (G ) r • Fact. #cut edges deleted in Step 2 is at most edges ( S , S ) log r (G ) min{d ( S ), d ( S )} min{d ( S ), d ( S )} log r k OPT r • Corollary. #edges deleted in total is at most k log1.5 r OPT Proof of k OPT • Lemma. (G ) r • Consider H = G \ OPT • For every (si, ti) pair, mincutH(si, ti) = |edges(Si, Ti)| < k (a witness cut) Si • Claim. The witness cuts are laminar si ti Ti Proof of Claim: witness cuts are laminar • Gomory-Hu Tree. (exists for every graph) A weighted tree that consists of edges representing all pairs minimum s-t cuts in the graph. mincutH(s, t) = mincutT(s, t) • All s-t mincuts in the tree are laminar ==> All mincuts in H are laminar ==> All witness cuts are laminar H: Gomory-Hu tree T Proof of k OPT • Lemma. (G ) r • Consider H = G \ OPT • For every (si, ti) pair, mincutH(si, ti) = |edges(Si, Ti)| < k (a witness cut) S2 • Claim. The witness cuts are laminar S1 • Let S1, S2, ..., Sm be the maximal witness cuts S3 Proof of k OPT • Lemma. (G ) r • Let S1, S2, ..., Sm be the maximal witness cuts in H=G\OPT S2 1. d(S1) + d(S2) + ... + d(Sm) >= r 2. edges OPT ( S i , S i ) 1 edges H ( S i , S i ) k 1 therefore S1 edges G ( S i , S i ) k edges OPT ( S i , S i ) S3 k OPT • Lemma. (G ) r Proof of • Let S1, S2, ..., Sm be the maximal witness cuts in H=G\OPT S2 1. d(S1) + d(S2) + ... + d(Sm) >= r 2. edges G ( S i , S i ) k edges OPT ( S i , S i ) m edges (S , S ) 2k OPT i 1 G i i (since each edge is shared by at most 2 maximal cuts) S1 S3 k OPT • Lemma. (G ) r Proof of • Let S1, S2, ..., Sm be the maximal witness cuts in H=G\OPT S2 1. d(S1) + d(S2) + ... + d(Sm) >= r 2. edges G ( S i , S i ) k edges OPT ( S i , S i ) m edges (S , S ) 2k OPT G i 1 i S1 i 3. by expansion m m edges (S , S ) (G)d (S ) i 1 G i i i 1 r (G) In all: r (G) 2k OPT i S3 (2, log2.5 r loglog r)-bicriteria approx. for general EC-kRC (sketch) (2, log2.5 r loglog r)-bicriteria approx. for general EC-kRC (sketch) • k-route non-uniform sparsity (k) wt (S , S ) (k ) (S ) , D( S , S ) S where : total wt of all the edges across the cut wt (S , S ) but the most expensive (k-1) ones (k) D(S , S ) : #source-sink pairs across the cut • Corollary. (of [ALN05]) O(log0.5 r loglog r) approx. in nO(k) time • Lemma. ( 2 k 1) ( k ) (G ) min ( k ) ( S ) OPT (G ) log r r (2, log2.5 r loglog r)-bicriteria approx. for general EC-kRC (sketch) (cont'd) • Lemma. ( 2 k 1) OPT (G ) log r r • The iterative algorithm. (Applying Idea 2) • Step 1. Use the algorithm in [ALN05] to find an approximate sparse cut • Step 2. Delete all the edges across the cut but the (2k2) most expensive ones • Step 3. Remove all the source-sink pairs that are (2k-1)disconnected • Step 4. Repeat Step 1~3 until no source-sink pair remains • Theorem. Wt. of removed edges <= log2.5 r loglog r OPT Open questions • Algorithm side. – Better true approximation algorithm for general ECkRC (and VC-kRC) • Hardness side. – Is EC-kRC (for large k) strictly harder than multicut? – Understand the simplest case: st-EC-kRC. Thank you!