Sparsest cut on bounded treewidth graphs: algorithms and hardness results

We give a 2-approximation algorithm for the non-uniform Sparsest Cut problem that runs in time nO(k), where k is the treewidth of the graph. This improves on the previous 22k-approximation in time poly(n) 2O(k) due to Chlamtac et al. [18]. To complement this algorithm, we show the following hardness results: If the non-uniform Sparsest Cut has a ρ-approximation for series-parallel graphs (where ρ ≥ 1), then the MaxCut problem has an algorithm with approximation factor arbitrarily close to 1/ρ. Hence, even for such restricted graphs (which have treewidth 2), the Sparsest Cut problem is NP-hard to approximate better than 17/16 - ε for ε > 0; assuming the Unique Games Conjecture the hardness becomes 1/αGW - ε. For graphs with large (but constant) treewidth, we show a hardness result of 2 - ε assuming the Unique Games Conjecture. Our algorithm rounds a linear program based on (a subset of) the Sherali-Adams lift of the standard Sparsest Cut LP. We show that even for treewidth-2 graphs, the LP has an integrality gap close to 2 even after polynomially many rounds of Sherali-Adams. Hence our approach cannot be improved even on such restricted graphs without using a stronger relaxation.

[1]  Venkatesan Guruswami,et al.  Constant Factor Lasserre Integrality Gaps for Graph Partitioning Problems , 2014, SIAM J. Optim..

[2]  Avner Magen,et al.  Robust Algorithms for on Minor-Free Graphs Based on the Sherali-Adams Hierarchy , 2009, APPROX-RANDOM.

[3]  David R. Karger,et al.  (De)randomized Construction of Small Sample Spaces in NC , 1997, J. Comput. Syst. Sci..

[4]  Venkatesan Guruswami,et al.  Certifying Graph Expansion and Non-Uniform Sparsity via Generalized Spectra , 2011, ArXiv.

[5]  Yuval Rabani,et al.  Improved lower bounds for embeddings into L1 , 2006, SODA '06.

[6]  B. Mohar,et al.  Graph Minors , 2009 .

[7]  Ehud Friedgut,et al.  Boolean Functions With Low Average Sensitivity Depend On Few Coordinates , 1998, Comb..

[8]  Eyal Amir,et al.  Approximation Algorithms for Treewidth , 2010, Algorithmica.

[9]  Nisheeth K. Vishnoi,et al.  The Unique Games Conjecture, Integrality Gap for Cut Problems and Embeddability of Negative Type Metrics into l1 , 2005, FOCS.

[10]  Prasad Raghavendra,et al.  Coarse Differentiation and Multi-flows in Planar Graphs , 2007, APPROX-RANDOM.

[11]  Satish Rao,et al.  Small distortion and volume preserving embeddings for planar and Euclidean metrics , 1999, SCG '99.

[12]  Anupam Gupta,et al.  Embedding k-outerplanar graphs into ℓ1 , 2003, SODA '03.

[13]  Johan Håstad,et al.  Some optimal inapproximability results , 2001, JACM.

[14]  Sanjeev Khanna,et al.  Polynomial flow-cut gaps and hardness of directed cut problems , 2007, STOC '07.

[15]  Moses Charikar,et al.  Integrality gaps for Sherali-Adams relaxations , 2009, STOC '09.

[16]  Subhash Khot,et al.  SDP Integrality Gaps with Local ell_1-Embeddability , 2009, 2009 50th Annual IEEE Symposium on Foundations of Computer Science.

[17]  Yuri Ilan,et al.  A Lower Bound on the Distortion of Embedding Planar Metrics into Euclidean Space , 2003, Discret. Comput. Geom..

[18]  Assaf Naor,et al.  . 20 24 v 2 [ cs . D S ] 18 N ov 2 00 9 A ( log n ) Ω ( 1 ) integrality gap for the Sparsest Cut SDP , 2009 .

[19]  Refael Hassin,et al.  On multicommodity flows in planar graphs , 1984, Networks.

[20]  Luca Trevisan,et al.  On Weighted vs Unweighted Versions of Combinatorial Optimization Problems , 2001, Inf. Comput..

[21]  Rishi Saket,et al.  SDP Integrality Gaps with Local̀ 1-Embeddability Subhash Khot , 2009 .

[22]  Farhad Shahrokhi,et al.  Sparsest cuts and bottlenecks in graphs , 1990, Discret. Appl. Math..

[23]  Paul D. Seymour,et al.  Graph Minors. II. Algorithmic Aspects of Tree-Width , 1986, J. Algorithms.

[24]  James R. Lee,et al.  Euclidean distortion and the sparsest cut , 2005, STOC '05.

[25]  Assaf Naor,et al.  Compression bounds for Lipschitz maps from the Heisenberg group to L1 , 2009, ArXiv.

[26]  Ryan O'Donnell,et al.  Noise stability of functions with low influences: Invariance and optimality , 2005, 46th Annual IEEE Symposium on Foundations of Computer Science (FOCS'05).

[27]  Prasad Raghavendra,et al.  Integrality Gaps for Strong SDP Relaxations of UNIQUE GAMES , 2009, 2009 50th Annual IEEE Symposium on Foundations of Computer Science.

[28]  James R. Lee,et al.  On the geometry of graphs with a forbidden minor , 2009, STOC '09.

[29]  Daniel Bienstock,et al.  Tree-width and the Sherali-Adams operator , 2004, Discret. Optim..

[30]  Luca Trevisan,et al.  Gadgets, Approximation, and Linear Programming , 2000, SIAM J. Comput..

[31]  Anupam Gupta,et al.  Cuts, Trees and ℓ1-Embeddings of Graphs* , 2004, Comb..

[32]  Satish Rao,et al.  Expander flows, geometric embeddings and graph partitioning , 2004, STOC '04.

[33]  Guy Kindler,et al.  Optimal inapproximability results for MAX-CUT and other 2-variable CSPs? , 2004, 45th Annual IEEE Symposium on Foundations of Computer Science.

[34]  James R. Lee,et al.  Near-optimal distortion bounds for embedding doubling spaces into L1 , 2011, STOC '11.

[35]  Hans L. Bodlaender,et al.  NC-Algorithms for Graphs with Small Treewidth , 1988, WG.

[36]  Ola Svensson,et al.  Inapproximability Results for Maximum Edge Biclique, Minimum Linear Arrangement, and Sparsest Cut , 2011, SIAM J. Comput..

[37]  Avner Magen,et al.  Robust Algorithms for Max Independent Set on Minor-Free Graphs Based on the Sherali-Adams Hierarchy , 2009 .

[38]  Yuri Rabinovich.  On average distortion of embedding metrics into the line and into L1 , 2003, STOC '03.

[39]  James R. Lee,et al.  Lp metrics on the Heisenberg group and the Goemans-Linial conjecture , 2006, 2006 47th Annual IEEE Symposium on Foundations of Computer Science (FOCS'06).

[40]  James R. Lee,et al.  Embeddings of Topological Graphs: Lossy Invariants, Linearization, and 2-Sums , 2008, 2008 49th Annual IEEE Symposium on Foundations of Computer Science.

[41]  Hans L. Bodlaender,et al.  A Partial k-Arboretum of Graphs with Bounded Treewidth , 1998, Theor. Comput. Sci..

[42]  Kenneth Ward Church,et al.  Nonlinear Estimators and Tail Bounds for Dimension Reduction in l1 Using Cauchy Random Projections , 2006, J. Mach. Learn. Res..

[43]  Oded Regev,et al.  Entropy-based bounds on dimension reduction in L1 , 2011 .

[44]  Michael I. Jordan,et al.  Treewidth-based conditions for exactness of the Sherali-Adams and Lasserre relaxations , 2004 .

[45]  Philip N. Klein,et al.  Excluded minors, network decomposition, and multicommodity flow , 1993, STOC.

[46]  Hans L. Bodlaender.  A linear time algorithm for finding tree-decompositions of small treewidth , 1993, STOC '93.

[47]  Prasad Raghavendra,et al.  Approximating Sparsest Cut in Graphs of Bounded Treewidth , 2010, APPROX-RANDOM.

[48]  Derek G. Corneil,et al.  Complexity of finding embeddings in a k -tree , 1987 .

[49]  Bruce A. Reed,et al.  Finding approximate separators and computing tree width quickly , 1992, STOC '92.

[50]  Chandra Chekuri,et al.  Flow-cut gaps for integer and fractional multiflows , 2010, SODA '10.

[51]  Assaf Naor,et al.  A $(\log n)^{\Omega(1)}$ Integrality Gap for the Sparsest Cut SDP , 2009, 2009 50th Annual IEEE Symposium on Foundations of Computer Science.

[52]  Yuval Rabani,et al.  ON THE HARDNESS OF APPROXIMATING MULTICUT AND SPARSEST-CUT , 2005, 20th Annual IEEE Conference on Computational Complexity (CCC'05).

[53]  J. R. Lee,et al.  Embedding the diamond graph in Lp and dimension reduction in L1 , 2004, math/0407520.

[54]  Y. Rabani,et al.  Improved lower bounds for embeddings into L 1 , 2006, SODA 2006.