The submodular partitioning problem asks to minimize, over all partitions P of a ground set V , the sum of a given submodular function f over the parts of P . The problem has seen considerable work in approximability, as it encompasses multiterminal cuts on graphs, k -cuts on hypergraphs, and elementary linear algebra problems such as matrix multiway partitioning. This research has been divided between the fixed terminal setting, where we are given a set of terminals that must be separated by P , and the global setting, where the only constraint is the size of the partition. We investigate a generalization that unifies these two settings: minimum submodular matroid-constrained partition. In this problem, we are additionally given a matroid over the ground set and seek to find a partition P in which there exists some basis that is separated by P . We explore the approximability of this problem and its variants for general, symmetric, and monotone submodular functions.

We consider deletion problems in graphs and supermodular functions where the goal is to reduce density. In Graph Density Deletion (GraphDD), we are given a graph G = (V,E) with non-negative vertex costs and a non-negative parameter ρ ≥ 0 and the goal is to remove a minimum cost subset S of vertices such that the densest subgraph in G-S has density at most ρ. This problem has an underlying matroidal structure and generalizes several classical problems such as vertex cover, feedback vertex set, and pseudoforest deletion set for appropriately chosen ρ ≤ 1 and all of these classical problems admit a 2-approximation. In sharp contrast, we prove that for every fixed integer ρ > 1, GraphDD is hard to approximate to within a logarithmic factor via a reduction from SetCover, thus showing a phase transition phenomenon. Next, we investigate a generalization of GraphDD to monotone supermodular functions, termed Supermodular Density Deletion (SupmodDD). In SupmodDD, we are given a monotone supermodular function f:2^V → ℤ_{≥0} via an evaluation oracle with element costs and a non-negative integer ρ ≥ 0 and the goal is remove a minimum cost subset S ⊆ V such that the densest subset according to f in V-S has density at most ρ. We show that SupmodDD is approximation equivalent to the well-known Submodular Cover problem; this implies a tight logarithmic approximation and hardness for SupmodDD; it also implies a logarithmic approximation for GraphDD, thus matching our inapproximability bound. Motivated by these hardness results, we design bicriteria approximation algorithms for both GraphDD and SupmodDD.

Finding the maximum number of disjoint spanning trees in a given graph is a well-studied problem with several applications and connections. The Tutte-Nash-Williams theorem provides a min-max relation for this problem which also extends to disjoint bases in a matroid and leads to efficient algorithms [Schrijver, 2003]. Several other packing problems such as element disjoint Steiner trees, disjoint set covers, and disjoint dominating sets are NP-Hard but admit an O(log n)-approximation [Feige et al., 2002; Cheriyan and Salavatipour, 2007]. Călinescu, Chekuri, and Vondrák [G. Călinescu et al., 2009] viewed all these packing problems as packing bases of a polymatroid and provided a unified perspective. Motivated by applications in wireless networks, recent works have studied the problem of packing set covers in the online model [Pananjady et al., 2015; Emek et al., 2019; Bienkowski et al., 2025]. The online model poses new challenges for packing problems. In particular, it is not clear how to pack a maximum number of disjoint spanning trees in a graph when edges arrive online. Motivated by these applications and theoretical considerations, we formulate an online model for packing bases of a polymatroid, and describe a randomized algorithm with a polylogarithmic competitive ratio. Our algorithm is based on interesting connections to the notion of quotients of a polymatroid that has recently seen applications in polymatroid sparsification [Quanrud, 2024]. We generalize the previously known result for the online disjoint set cover problem [Emek et al., 2019] and also address several other packing problems in a unified fashion. For the special case of packing disjoint spanning trees in a graph (or a hypergraph) whose edges arrive online, we provide an alternative to our general algorithm that is simpler and faster while achieving the same poly-logarithmic competitive ratio.

Scarf’s algorithm - a pivoting procedure that finds a dominating extreme point in a down-monotone polytope - can be used to show the existence of a fractional stable matching in hypergraphs. The problem of finding a fractional stable matching in hypergraphs, however, is PPAD-complete. In this work, we study the behavior of Scarf’s algorithm on arborescence hypergraphs, the family of hypergraphs in which hyperedges correspond to the paths of an arborescence. For arborescence hypergraphs, we prove that Scarf’s algorithm can be implemented to find an integral stable matching in polynomial time. En route to our result, we uncover novel structural properties of bases and pivots for the more general family of network hypergraphs. Our work provides the first proof of polynomial-time convergence of Scarf’s algorithm on hypergraphic stable matching problems, giving hope to the possibility of polynomial-time convergence of Scarf’s algorithm for other families of polytopes.

Flows and colorings are disparate concepts in graph algorithms -- the former is tractable while the latter is intractable. Tutte introduced the concept of nowhere-zero flows to unify these two concepts. Jaeger showed that nowhere-zero flows are equivalent to cut-balanced orientations. Motivated by connections between nowhere-zero flows, cut-balanced orientations, Nash-Williams' well-balanced orientations, and postman problems, we study optimization versions of nowhere-zero flows and cut-balanced orientations. Given a bidirected graph with asymmetric costs on two orientations of each edge, we study the min cost nowhere-zero $k$-flow problem and min cost $k$-cut-balanced orientation problem. We show that both problems are NP-hard to approximate within any finite factor. Given the strong inapproximability result, we design bicriteria approximations for both problems: we obtain a $(6,6)$-approximation to the min cost nowhere-zero $k$-flow and a $(k,6)$-approximation to the min cost $k$-cut-balanced orientation. For the case of symmetric costs (where the costs of both orientations are the same for every edge), we show that the nowhere-zero $k$-flow problem remains NP-hard and admits a $3$-approximation.

The splitting-off operation in undirected graphs is a fundamental reduction operation that detaches all edges incident to a given vertex and adds new edges between the neighbors of that vertex while preserving their degrees. Lovász [Lov{á}sz, 1974; Lov{á}sz, 1993] and Mader [Mader, 1978] showed the existence of this operation while preserving global and local connectivities respectively in graphs under certain conditions. These results have far-reaching applications in graph algorithms literature [Lovász, 1976; Mader, 1978; Frank, 1993; Frank and Király, 2002; Király and Lau, 2008; Frank, 1992; Goemans and Bertsimas, 1993; Frank, 1994; Bang-Jensen et al., 1995; Frank, 2011; Nagamochi and Ibaraki, 2008; Nagamochi et al., 1997; Henzinger and Williamson, 1996; Goemans, 2001; Jordán, 2003; Kriesell, 2003; Jain et al., 2003; Chan et al., 2011; Bhalgat et al., 2008; Lau, 2007; Chekuri and Shepherd, 2008; Nägele and Zenklusen, 2020; Blauth and Nägele, 2023]. In this work, we introduce a splitting-off operation in hypergraphs. We show that there exists a local connectivity preserving complete splitting-off in hypergraphs and give a strongly polynomial-time algorithm to compute it in weighted hypergraphs. We illustrate the usefulness of our splitting-off operation in hypergraphs by showing two applications: (1) we give a constructive characterization of k-hyperedge-connected hypergraphs and (2) we give an alternate proof of an approximate min-max relation for max Steiner rooted-connected orientation of graphs and hypergraphs (due to Király and Lau [Király and Lau, 2008]). Our proof of the approximate min-max relation for graphs circumvents the Nash-Williams' strong orientation theorem and uses tools developed for hypergraphs.

Abstract. In submodular [Formula: see text]-partition, the input is a submodular function [Formula: see text] (given by an evaluation oracle) along with a positive integer [Formula: see text], and the goal is to find a partition of the ground set [Formula: see text] into [Formula: see text] nonempty parts [Formula: see text] in order to minimize [Formula: see text]. Narayanan, Roy, and Patkar [ J. Algorithms, 21 (1996), pp. 306–330] designed an algorithm for submodular [Formula: see text]-partition based on the principal partition sequence and showed that the approximation factor of their algorithm is 2 for the special case of graph cut functions (which was subsequently rediscovered by Ravi and Sinha [ European J. Oper. Res., 186 (2008), pp. 77–90]). In this work, we study the approximation factor of their algorithm for three subfamilies of submodular functions—namely monotone, symmetric, and posimodular—and show the following results: (1) The approximation factor of their algorithm for monotone submodular [Formula: see text]-partition is [Formula: see text]. This result improves on the 2-factor that was known to be achievable for monotone submodular [Formula: see text]-partition via other algorithms. Moreover, our upper bound of [Formula: see text] matches the recently shown lower bound under polynomial number of function evaluation queries [R. Santiago, Proceedings of the International Workshop on Combinatorial Algorithms, IWOCA, 2021, pp. 516–530]. Our upper bound of [Formula: see text] is also the first improvement beyond 2 for a certain graph partitioning problem that is a special case of monotone submodular [Formula: see text]-partition. (2) The approximation factor of their algorithm for symmetric submodular [Formula: see text]-partition is 2. This result generalizes their approximation factor analysis beyond graph cut functions. (3) The approximation factor of their algorithm for posimodular submodular [Formula: see text]-partition is 2. We also construct an example to show that the approximation factor of their algorithm for arbitrary submodular functions is [Formula: see text].

In submodular k-partition, the input is a submodular function f:2^V → ℝ_{≥ 0} (given by an evaluation oracle) along with a positive integer k and the goal is to find a partition of the ground set V into k non-empty parts V_1, V_2, …, V_k in order to minimize ∑_{i=1}^k f(V_i). Narayanan, Roy, and Patkar [Narayanan et al., 1996] designed an algorithm for submodular k-partition based on the principal partition sequence and showed that the approximation factor of their algorithm is 2 for the special case of graph cut functions (which was subsequently rediscovered by Ravi and Sinha [R. Ravi and A. Sinha, 2008]). In this work, we study the approximation factor of their algorithm for three subfamilies of submodular functions - namely monotone, symmetric, and posimodular and show the following results: 1) The approximation factor of their algorithm for monotone submodular k-partition is 4/3. This result improves on the 2-factor that was known to be achievable for monotone submodular k-partition via other algorithms. Moreover, our upper bound of 4/3 matches the recently shown lower bound under polynomial number of function evaluation queries [Santiago, 2021]. Our upper bound of 4/3 is also the first improvement beyond 2 for a certain graph partitioning problem that is a special case of monotone submodular k-partition. 2) The approximation factor of their algorithm for symmetric submodular k-partition is 2. This result generalizes their approximation factor analysis beyond graph cut functions. 3) The approximation factor of their algorithm for posimodular submodular k-partition is 2. We also construct an example to show that the approximation factor of their algorithm for arbitrary submodular functions is Ω(n/k).

We consider the problem of enumerating optimal solutions for two hypergraph $k$-partitioning problems -- namely, Hypergraph-$k$-Cut and Minmax-Hypergraph-$k$-Partition. The input in hypergraph $k$-partitioning problems is a hypergraph $G=(V, E)$ with positive hyperedge costs along with a fixed positive integer $k$. The goal is to find a partition of $V$ into $k$ non-empty parts $(V_1, V_2, \ldots, V_k)$ -- known as a $k$-partition -- so as to minimize an objective of interest. 1. If the objective of interest is the maximum cut value of the parts, then the problem is known as Minmax-Hypergraph-$k$-Partition. A subset of hyperedges is a minmax-$k$-cut-set if it is the subset of hyperedges crossing an optimum $k$-partition for Minmax-Hypergraph-$k$-Partition. 2. If the objective of interest is the total cost of hyperedges crossing the $k$-partition, then the problem is known as Hypergraph-$k$-Cut. A subset of hyperedges is a min-$k$-cut-set if it is the subset of hyperedges crossing an optimum $k$-partition for Hypergraph-$k$-Cut. We give the first polynomial bound on the number of minmax-$k$-cut-sets and a polynomial-time algorithm to enumerate all of them in hypergraphs for every fixed $k$. Our technique is strong enough to also enable an $n^{O(k)}p$-time deterministic algorithm to enumerate all min-$k$-cut-sets in hypergraphs, thus improving on the previously known $n^{O(k^2)}p$-time deterministic algorithm, where $n$ is the number of vertices and $p$ is the size of the hypergraph. The correctness analysis of our enumeration approach relies on a structural result that is a strong and unifying generalization of known structural results for Hypergraph-$k$-Cut and Minmax-Hypergraph-$k$-Partition. We believe that our structural result is likely to be of independent interest in the theory of hypergraphs (and graphs).

We consider the problem of deterministically enumerating all minimum k-cut-sets in a given hypergraph for any fixed k. The input here is a hypergraph G = (V, E) with non-negative hyperedge costs. A subset F ⊆ E of hyperedges is a k-cut-set if the number of connected components in G–F is at least k and it is a minimum k-cut-set if it has the least cost among all k-cut-sets. For fixed k, we call the problem of finding a minimum k-cut-set as Hypergraph-k-Cut and the problem of enumerating all minimum k-cut-sets as Enum-Hypergraph-k-Cut. The special cases of Hypergraph-k-Cut and Enum-Hypergraph-k-Cut restricted to graph inputs are well-known to be solvable in (randomized as well as deterministic) polynomial time [17,25,28,39]. In contrast, it is only recently that polynomial-time algorithms for Hypergraph-k-Cut were developed [2,3,12]. The randomized polynomial-time algorithm for Hypergraph-k-Cut that was designed in 2018 [3] showed that the number of minimum k-cut-sets in a hypergraph is O(n2k–2), where n is the number of vertices in the input hypergraph, and that they can all be enumerated in randomized polynomial time, thus resolving Enum-Hypergraph-k-Cut in randomized polynomial time. A deterministic polynomial-time algorithm for Hypergraph-k-Cut was subsequently designed in 2020 [2], but it is not guaranteed to enumerate all minimum k-cut-sets. In this work, we give the first deterministic polynomial-time algorithm to solve Enum-Hypergraph-k-Cut (this is non-trivial even for k = 2). Our algorithm is based on new structural results that allow for efficient recovery of all minimum k-cut-sets by solving minimum (S, T)-terminal cuts. Our techniques give new structural insights even for enumerating all minimum cut-sets (i.e., minimum 2-cut-sets) in a given hypergraph.

We introduce and study 𝓁_p-norm-multiway-cut: the input here is an undirected graph with non-negative edge weights along with k terminals and the goal is to find a partition of the vertex set into k parts each containing exactly one terminal so as to minimize the 𝓁_p-norm of the cut values of the parts. This is a unified generalization of min-sum multiway cut (when p = 1) and min-max multiway cut (when p = ∞), both of which are well-studied classic problems in the graph partitioning literature. We show that 𝓁_p-norm-multiway-cut is NP-hard for constant number of terminals and is NP-hard in planar graphs. On the algorithmic side, we design an O(log² n)-approximation for all p ≥ 1. We also show an integrality gap of Ω(k^{1-1/p}) for a natural convex program and an O(k^{1-1/p-ε})-inapproximability for any constant ε > 0 assuming the small set expansion hypothesis.

We consider the complexity of minmax partitioning of graphs, hypergraphs and (symmetric) submodular functions. Our main result is an algorithm for the problem of partitioning the ground set of a given symmetric submodular function f : 2V → ℝ into k non-empty parts V1, V2, …, Vk to minimize . Our algorithm runs in time, where n = |V| and T is the time to evaluate f on a given set; hence. this yields a polynomial time algorithm for any fixed k in the evaluation oracle model. As an immediate corollary, for any fixed k, there is a polynomial-time algorithm for the problem of partitioning the vertex set of a given hypergraph H = (V, E) into k non-empty parts to minimize the maximum capacity of the parts. The complexity of this problem. termed Minmax-Hypergraph-k-Part, was raised by Lawler in 1973 [16]. In contrast to our positive result, the reduction in [6] implies that when k is part of the input, Minmax-Hypergraph-k-Part is hard to approximate to within an almost polynomial factor under the Exponential Time Hypothesis (ETH).

Finding locally optimal solutions for MAX-CUT and MAX- k -CUT are well-known PLS-complete problems. An instinctive approach to finding such a locally optimum solution is the FLIP method. Even though FLIP requires exponential time in worst-case instances, it tends to terminate quickly in practical instances. To explain this discrepancy, the run-time of FLIP has been studied in the smoothed complexity framework. Etscheid and Röglin (ACM Transactions on Algorithms, 2017) showed that the smoothed complexity of FLIP for max-cut in arbitrary graphs is quasi-polynomial. Angel, Bubeck, Peres, and Wei (STOC, 2017) showed that the smoothed complexity of FLIP for max-cut in complete graphs is ( O Φ 5 n 15.1 ), where Φ is an upper bound on the random edge-weight density and Φ is the number of vertices in the input graph. While Angel, Bubeck, Peres, and Wei’s result showed the first polynomial smoothed complexity, they also conjectured that their run-time bound is far from optimal. In this work, we make substantial progress toward improving the run-time bound. We prove that the smoothed complexity of FLIP for max-cut in complete graphs is O (Φ n 7.83 ). Our results are based on a carefully chosen matrix whose rank captures the run-time of the method along with improved rank bounds for this matrix and an improved union bound based on this matrix. In addition, our techniques provide a general framework for analyzing FLIP in the smoothed framework. We illustrate this general framework by showing that the smoothed complexity of FLIP for MAX-3-CUT in complete graphs is polynomial and for MAX - k - CUT in arbitrary graphs is quasi-polynomial. We believe that our techniques should also be of interest toward showing smoothed polynomial complexity of FLIP for MAX - k - CUT in complete graphs for larger constants k .

We address counting and optimization variants of multicriteria global min-cut and size-constrained min-$k$-cut in hypergraphs. 1. For an $r$-rank $n$-vertex hypergraph endowed with $t$ hyperedge-cost functions, we show that the number of multiobjective min-cuts is $O(r2^{tr}n^{3t-1})$. In particular, this shows that the number of parametric min-cuts in constant rank hypergraphs for a constant number of criteria is strongly polynomial, thus resolving an open question by Aissi, Mahjoub, McCormick, and Queyranne (Math Programming, 2015). In addition, we give randomized algorithms to enumerate all multiobjective min-cuts and all pareto-optimal cuts in strongly polynomial-time. 2. We also address node-budgeted multiobjective min-cuts: For an $n$-vertex hypergraph endowed with $t$ vertex-weight functions, we show that the number of node-budgeted multiobjective min-cuts is $O(r2^{r}n^{t+2})$, where $r$ is the rank of the hypergraph, and the number of node-budgeted $b$-multiobjective min-cuts for a fixed budget-vector $b$ is $O(n^2)$. 3. We show that min-$k$-cut in hypergraphs subject to constant lower bounds on part sizes is solvable in polynomial-time for constant $k$, thus resolving an open problem posed by Queyranne. Our technique also shows that the number of optimal solutions is polynomial. All of our results build on the random contraction approach of Karger (SODA, 1993). Our techniques illustrate the versatility of the random contraction approach to address counting and algorithmic problems concerning multiobjective min-cuts and size-constrained $k$-cuts in hypergraphs.

A heapable sequence is a sequence of numbers that can be arranged in a min-heap data structure. Finding a longest heapable subsequence of a given sequence was proposed by Byers, Heeringa, Mitzenmacher, and Zervas (ANALCO 2011) as a generalization of the well-studied longest increasing subsequence problem and its complexity still remains open. An equivalent formulation of the longest heapable subsequence problem is that of finding a maximum-sized binary tree in a given permutation directed acyclic graph (permutation DAG). In this work, we study parameterized algorithms for both longest heapable subsequence and maximum-sized binary tree. We introduce alphabet size as a new parameter in the study of computational problems in permutation DAGs and show that this parameter with respect to a fixed topological ordering admits a complete characterization and a polynomial time algorithm. We believe that this parameter is likely to be useful in the context of optimization problems defined over permutation DAGs.

We consider the Hypergraph- k-Cut problem. The input consists of a hypergraph G = (V, E) with nonnegative hyperedge-costs c:E→ \mathbbR+ and a positive integer k. The objective is to find a least-cost subset F ⊆ E such that the number of connected components in G-F is at least k. An alternative formulation of the objective is to find a partition of V into k non-empty sets V1, V2, ..., Vk so as to minimize the cost of the hyperedges that cross the partition. Graph- k-Cut, the special case of Hypergraph- k-Cut obtained by restricting to graph inputs, has received considerable attention. Several different approaches lead to a polynomial-time algorithm for Graph- k-Cut when k is fixed, starting with the work of Goldschmidt and Hochbaum (1988) [1], [2]. In contrast, it is only recently that a randomized polynomial time algorithm for Hypergraph- k-Cut was developed [3] via a subtle generalization of Karger's random contraction approach for graphs. In this work, we develop the first deterministic polynomial time algorithm for Hypergraph- k-Cut for all fixed k. We describe two algorithms both of which are based on a divide and conquer approach. The first algorithm is simpler and runs in nO(k2) time while the second one runs in nO(k)time. Our proof relies on new structural results that allow for efficient recovery of the parts of an optimum k-partition by solving minimum ( S, T) -terminal cuts. Our techniques give new insights even for Graph- k-Cut.

A heapable sequence is a sequence of numbers that can be arranged in a min-heap data structure. Finding a longest heapable subsequence of a given sequence was proposed by Byers, Heeringa, Mitzenmacher, and Zervas (ANALCO 2011) as a generalization of the well-studied longest increasing subsequence problem and its complexity still remains open. An equivalent formulation of the longest heapable subsequence problem is that of finding a maximum-sized binary tree in a given permutation directed acyclic graph (permutation DAG). In this work, we study parameterized algorithms for both longest heapable subsequence and maximum-sized binary tree. We introduce alphabet size as a new parameter in the study of computational problems in permutation DAGs and show that this parameter with respect to a fixed topological ordering admits a complete characterization and a polynomial time algorithm. We believe that this parameter is likely to be useful in the context of optimization problems defined over permutation DAGs.

We investigate the odd multiway node (edge) cut problem where the input is a graph with a specified collection of terminal nodes, and the goal is to find a smallest subset of non-terminal nodes (edges) to delete so that the terminal nodes do not have an odd length path between them. In an earlier work, Lokshtanov and Ramanujan showed that both odd multiway node cut and odd multiway edge cut are fixed-parameter tractable (FPT) when parameterized by the size of the solution in undirected graphs. In this work, we focus on directed acyclic graphs (DAGs) and design a fixed-parameter algorithm. Our main contribution is a broadening of the shadow-removal framework to address parity problems in DAGs. We complement our FPT results with tight approximability as well as polyhedral results for two terminals in DAGs. Additionally, we show inapproximability results for odd multiway edge cut in undirected graphs even for two terminals.

The spectra of signed matrices have played a fundamental role in social sciences, graph theory, and control theory. In this work, we investigate the computational problems of finding symmetric signings of matrices with natural spectral properties. Our results are the following: 1) We characterize matrices that have an invertible signing: a symmetric matrix has an invertible symmetric signing if and only if the support graph of the matrix contains a perfect 2-matching. Further, we present an efficient algorithm to search for an invertible symmetric signing. 2) We use the above-mentioned characterization to give an algorithm to find a minimum increase in the support of a given symmetric matrix so that it has an invertible symmetric signing. 3) We show NP-completeness of the following problems: verifying whether a given matrix has a symmetric signing that is singular or has bounded eigenvalues. However, we also illustrate that the complexity could differ substantially for input matrices that are adjacency matrices of graphs. We use combinatorial techniques in addition to classic results from matching theory.

Finding locally optimal solutions for MAX-CUT and MAX-k-CUT are well-known PLS-complete problems. An instinctive approach to finding such a locally optimum solution is the FLIP method. Even though FLIP requires exponential time in worst-case instances, it tends to terminate quickly in practical instances. To explain this discrepancy, the run-time of FLIP has been studied in the smoothed complexity framework. Etscheid and Röglin [ER17] showed that the smoothed complexity of FLIP for MAX-CUT in arbitrary graphs is quasi-polynomial. Angel, Bubeck, Peres and Wei [ABPW17] showed that the smoothed complexity of FLIP for MAX-CUT in complete graphs is O(ϕ5n15.1), where ϕ is an upper bound on the random edge-weight density and n is the number of vertices in the input graph. While Angel et al.'s result showed the first polynomial smoothed complexity, they also conjectured that their run-time bound is far from optimal. In this work, we make substantial progress towards improving the run-time bound. We prove that the smoothed complexity of FLIP in complete graphs is O(ϕn7.83). Our results are based on a carefully chosen matrix whose rank captures the run-time of the method along with improved rank bounds for this matrix and an improved union bound based on this matrix. In addition, our techniques provide a general framework for analyzing FLIP in the smoothed framework. We illustrate this general framework by showing that the smoothed complexity of FLIP for MAX-3-CUT in complete graphs is polynomial and for MAX-k-CUT in arbitrary graphs is quasi-polynomial. We believe that our techniques should also be of interest towards showing smoothed polynomial complexity of FLIP for MAX-k-CUT in complete graphs for larger constants k.

A $k$-lift of an $n$-vertex base graph $G$ is a graph $H$ on $n\times k$ vertices, where each vertex $v$ of $G$ is replaced by $k$ vertices $v_1,\ldots,v_k$ and each edge $uv$ in $G$ is replaced by a matching representing a bijection $\pi_{uv}$ so that the edges of $H$ are of the form $(u_i,v_{\pi_{uv}(i)})$. Lifts have been investigated as a means to efficiently construct expanders. In this work, we study lifts obtained from groups and group actions. We derive the spectrum of such lifts via the representation theory principles of the underlying group. Our main results are 1. a uniform random lift by a cyclic group of order $k$ of any $n$-vertex $d$-regular base graph $G$, with the nontrivial eigenvalues of the adjacency matrix of $G$ bounded by $\lambda$ in magnitude, has the new nontrivial eigenvalues bounded by $\lambda+\mathcal{O}(\sqrt{d})$ in magnitude with probability $1-ke^{-\Omega(n/d^2)}$. The probability bounds as well as the dependency on $\lambda$ are almost optimal. As a special case, we obtain that there is a constant $c_1$ such that for every $k\leq 2^{c_1n/d^2}$, there exists a lift $H$ of every Ramanujan graph by a cyclic group of order $k$ such that $H$ is almost Ramanujan (nontrivial eigenvalues of the adjacency matrix at most $O(\sqrt{d})$ in magnitude). This result leads to a quasi-polynomial time deterministic algorithm to construct almost Ramanujan expanders; 2. there is a constant $c_2$ such that for every $k\geq 2^{c_2nd}$, there does not exist an abelian $k$-lift $H$ of any $n$-vertex $d$-regular base graph such that $H$ is almost Ramanujan. This can be viewed as an analogue of the well-known nonexpansion result for constant degree abelian Cayley graphs. Suppose $k_0$ is the order of the largest abelian group that produces expanding lifts. Our two results highlight lower and upper bounds on $k_0$ that are tight up to a factor of $d^3$ in the exponent, thus suggesting a threshold phenomenon.

Locality sensitive hashing (LSH) was introduced by Indyk and Motwani (STOC'98) to give the first sublinear time algorithm for the c-approximate nearest neighbor (ANN) problem using only polynomial space. At a high level, an LSH family hashes "nearby" points to the same bucket and "far away" points to different buckets. The quality of measure of an LSH family is its LSH exponent, which helps determine both query time and space usage. In a seminal work, Andoni and Indyk (FOCS '06) constructed an LSH family based on random ball partitionings of space that achieves an LSH exponent of 1/c^2 for the l_2 norm, which was later shown to be optimal by Motwani, Naor and Panigrahy (SIDMA '07) and O'Donnell, Wu and Zhou (TOCT '14). Although optimal in the LSH exponent, the ball partitioning approach is computationally expensive. So, in the same work, Andoni and Indyk proposed a simpler and more practical hashing scheme based on Euclidean lattices and provided computational results using the 24-dimensional Leech lattice. However, no theoretical analysis of the scheme was given, thus leaving open the question of finding the exponent of lattice based LSH. In this work, we resolve this question by showing the existence of lattices achieving the optimal LSH exponent of 1/c^2 using techniques from the geometry of numbers. At a more conceptual level, our results show that optimal LSH space partitions can have periodic structure. Understanding the extent to which additional structure can be imposed on these partitions, e.g. to yield low space and query complexity, remains an important open problem.

We investigate the approximability of the linear 3-cut problem in directed graphs, which is the simplest unsolved case of the linear k-cut problem. The input here is a directed graph D = (V, E) with node weights and three specified terminal nodes s,r,t ∊ V, and the goal is to find a minimum weight subset of non-terminal nodes whose removal ensures that s cannot reach r and t, and r cannot reach t. The problem is approximation-equivalent to the problem of blocking rooted in- and out-arborescences, and it also has applications in network coding and security. The approximability of linear 3-cut has been wide open until now: the best known lower bound under the Unique Games Conjecture (UGC) was 4/3, while the best known upper bound was 2 using a trivial algorithm. In this work we completely close this gap: we present a -approximation algorithm and show that this factor is tight assuming UGC. Our contributions are twofold: (1) we analyze a natural two-step deterministic rounding scheme through the lens of a single-step randomized rounding scheme with non-trivial distributions, and (2) we construct integrality gap instances that meet the upper bound of . Our gap instances can be viewed as a weighted graph sequence converging to a “graph limit structure”.

In the hypergraph k-cut problem, the input is a hypergraph, and the goal is to find a smallest subset of hyperedges whose removal ensures that the remaining hypergraph has at least k connected components. This problem is known to be at least as hard as the densest k-subgraph problem when k is part of the input (Chekuri-Li, 2015). We present a randomized polynomial time algorithm to solve the hypergraph k-cut problem for constant k. Our algorithm solves the more general hedge k-cut problem when the subgraph induced by every hedge has a constant number of connected components. In the hedge k-cut problem, the input is a hedgegraph specified by a vertex set and a disjoint set of hedges, where each hedge is a subset of edges defined over the vertices. The goal is to find a smallest subset of hedges whose removal ensures that the number of connected components in the remaining underlying (multi-)graph is at least k. Our algorithm is based on random contractions akin to Karger's min cut algorithm. Our main technical contribution is a distribution over the hedges (hyperedges) so that random contraction of hedges (hyperedges) chosen from the distribution succeeds in returning an optimum solution with large probability.

Testing membership in lattices is of practical relevance, with applications to integer programming, error detection in lattice-based communication, and cryptography. In this work, we initiate a systematic study of local testing for membership in lattices, complementing and building upon the extensive body of work on locally testable codes. In particular, we formally define the notion of local tests for lattices and present the following: 1. We show that in order to achieve low query complexity, it is sufficient to design $1$-sided nonadaptive canonical tests. This result is akin to, and based on, an analogous result for error-correcting codes due to [E. Ben-Sasson, P. Harsha, and S. Raskhodnikova, SIAM J. Comput., 35 (2005), pp. 1--21]. 2. We demonstrate upper and lower bounds on the query complexity of local testing for membership in code formula lattices. We instantiate our results for code formula lattices constructed from Reed--Muller codes to obtain nearly matching upper and lower bounds on the query complexity of testing such lattices. 3. We contrast lattice testing to code testing by showing lower bounds on the query complexity of testing low-dimensional lattices. This illustrates large lower bounds on the query complexity of testing membership in the well-known knapsack lattices. On the other hand, we show that knapsack lattices with bounded coefficients have low-query testers if the inputs are promised to lie in the span of the lattice.

The computational complexity of multicut-like problems may vary significantly depending on whether the terminals are fixed or not. In this work we present a comprehensive study of this phenomenon in two types of cut problems in directed graphs: double cut and bicut. 1. Fixed-terminal edge-weighted double cut is known to be solvable efficiently. We show that fixed-terminal node-weighted double cut cannot be approximated to a factor smaller than 2 under the Unique Games Conjecture (UGC), and we also give a 2-approximation algorithm. For the global version of the problem, we prove an inapproximability bound of 3/2 under UGC. 2. Fixed-terminal edge-weighted bicut is known to have an approximability factor of 2 that is tight under UGC. We show that the global edge-weighted bicut is approximable to a factor strictly better than 2, and that the global node-weighted bicut cannot be approximated to a factor smaller than 3/2 under UGC. 3. In relation to these investigations, we also prove two results on undirected graphs which are of independent interest. First, we show NP-completeness and a tight inapproximability bound of 4/3 for the node-weighted 3-cut problem under UGC. Second, we show that for constant k, there exists an efficient algorithm to solve the minimum {s,t}-separating k-cut problem. Our techniques for the algorithms are combinatorial, based on LPs and based on the enumeration of approximate min-cuts. Our hardness results are based on combinatorial reductions and integrality gap instances.

Lattices are discrete mathematical objects with widespread applications to integer programs as well as modern cryptography. An important class of lattices are those that possess an orthogonal basis, since if such an orthogonal basis is known, then many other fundamental problems on lattices can be solved easily (e.g., the Closest Vector Problem). However, intriguingly, deciding whether a lattice has an orthogonal basis is not known to be either NP-complete or in P. In this paper, we focus on the orthogonality decision problem for a well-known family of lattices, namely Construction-A lattices. These are lattices of the form $C+q\mathbb{Z}^n$, where $C$ is an error-correcting $q$-ary code, and are studied in communication settings. We provide a complete characterization of lattices obtained from binary and ternary codes using Construction-A that have an orthogonal basis. We use this characterization to give an efficient algorithm to solve the orthogonality decision problem. Our algorithm also finds an orthogonal basis if one exists for this family of lattices.

Testing membership in lattices is of practical relevance, with applications to integer programming, error detection in lattice-based communication and cryptography. In this work, we initiate a systematic study of local testing for membership in lattices, complementing and building upon the extensive body of work on locally testable codes. In particular, we formally define the notion of local tests for lattices and present the following: 1. We show that in order to achieve low query complexity, it is sufficient to design one-sided non-adaptive canonical tests. This result is akin to, and based on an analogous result for error-correcting codes due to Ben-Sasson et al. (SIAM J. Computing, 35(1):1-21). 2. We demonstrate upper and lower bounds on the query complexity of local testing for membership in code formula lattices. We instantiate our results for code formula lattices constructed from Reed-Muller codes to obtain nearly-matching upper and lower bounds on the query complexity of testing such lattices. 3. We contrast lattice testing from code testing by showing lower bounds on the query complexity of testing low-dimensional lattices. This illustrates large lower bounds on the query complexity of testing membership in knapsack lattices. On the other hand, we show that knapsack lattices with bounded coefficients have low-query testers if the inputs are promised to lie in the span of the lattice.

We study the problem of learning a most biased coin among a set of coins by tossing the coins adaptively. The goal is to minimize the number of tosses until we identify a coin whose posterior probability of being most biased is at least 1-δfor a given δ. Under a particular probabilistic model, we give an optimal algorithm, i.e., an algorithm that minimizes the expected number of future tosses. The problem is closely related to finding the best arm in the multi-armed bandit problem using adaptive strategies. Our algorithm employs an optimal adaptive strategy—a strategy that performs the best possible action at each step after observing the outcomes of all previous coin tosses. Consequently, our algorithm is also optimal for any given starting history of outcomes. To our knowledge, this is the first algorithm that employs an optimal adaptive strategy under a Bayesian setting for this problem. Our proof of optimality employs mathematical tools from the area of Markov games.

We study the problem of answering k-way marginal queries on a database D ϵ ({0,1}d)n, while preserving differential privacy. The answer to a k-way marginal query is the fraction of the database's records x in {0,1}d with a given value in each of a given set of up to k columns. Marginal queries enable a rich class of statistical analyses on a dataset, and designing efficient algorithms for privately answering marginal queries has been identified as an important open problem in private data analysis. For any k, we give a differentially private online algorithm that runs in time poly (n, 2o(d)) per query and answers any sequence of poly(n) many k-way marginal queries with error at most ±0.01 on every query, provided n ≥ d0.51. To the best of our knowledge, this is the first algorithm capable of privately answering marginal queries with a non-trivial worst-case accuracy guarantee for databases containing poly(d, k) records in time exp(o(d)). Our algorithm runs the private multiplicative weights algorithm (Hardt and Rothblum, FOCS '10) on a new approximate polynomial representation of the database.

We study the Chance-Constrained Integer Feasibility Problem, where the goal is to determine whether the random polytope P(A,b)={x ϵ Rn : Aix ≤ bi, i ϵ [m]} obtained by choosing the constraint matrix A and vector b from a known distribution is integer feasible with probability at least 1-ε. We consider the case when the entries of the constraint matrix A are i.i.d. Gaussian (equivalently are i.i.d. from any spherically symmetric distribution). The radius of the largest inscribed ball is closely related to the existence of integer points in the polytope. We find that for m up to 2O(√n) constraints (rows of A), there exist constants c0 < c1 such that with high probability (ɛ = 1 /poly(n)), random polytopes are integer feasible if the radius of the largest ball contained in the polytope is at least c1√log(m/n)); and integer infeasible if the largest ball contained in the polytope is centered at (1/2,...,1/2) and has radius at most c0√log(m/n)). Thus, random polytopes transition from having no integer points to being integer feasible within a constant factor increase in the radius of the largest inscribed ball. Integer feasibility is based on a randomized polynomial-time algorithm for finding an integer point in the polytope.

The Lovász local lemma (LLL) [P. Erdös and L. Lovász, Problems and results on 3-chromatic hypergraphs and some related questions, in Infinite and Finite Sets, Vol. II, A. Hajnal, R. Rado, and V. T. Sós, eds., North--Holland, Amsterdam, 1975, pp. 609--627] is a powerful result in probability theory that informally states the following: the probability that none of a set of bad events happens is positive if the probability of each event is small compared to the number of events that depend on it. The LLL is often used for nonconstructive existence proofs of combinatorial structures. A prominent application is to $k$-CNF formulas, where the LLL implies that if every clause in a formula shares variables with at most $d \leq 2^k/e-1$ other clauses, then such a formula has a satisfying assignment. Recently, a randomized algorithm to efficiently construct a satisfying assignment in this setting was given by Moser [A constructive proof of the Lovász local lemma, in STOC '09: Proceedings of the 41st Annual ACM Symposium on Theory of Computing, ACM, New York, 2009, pp. 343--350]. Subsequently Moser and Tardos [J. ACM, 57 (2010), pp. 11:1--11:15] gave a general algorithmic framework for the LLL and a randomized algorithm within this framework to construct the structures guaranteed by the LLL. The main problem left open by Moser and Tardos was to design an efficient deterministic algorithm for constructing structures guaranteed by the LLL. In this paper we provide such an algorithm. Our algorithm works in the general framework of Moser and Tardos with a minimal loss in parameters. For the special case of constructing satisfying assignments for $k$-CNF formulas with $m$ clauses, where each clause shares variables with at most $d \leq 2^{k/(1+\epsilon)}/e - 1$ other clauses, for any $\epsilon\in (0,1)$, we give a deterministic algorithm that finds a satisfying assignment in time $\tilde{O}(m^{2(1+1/\epsilon)})$. This improves upon the deterministic algorithms of Moser and of Moser and Tardos with running times $m^{\Omega(k^2)}$ and $m^{\Omega(d \log d)}$, respectively, which are superpolynomial for $k=\omega(1)$ and $d=\omega(1)$, and upon the previous best deterministic algorithm of Beck, which runs in polynomial time only for $d\leq 2^{k/16}/4$. Our algorithm is the first deterministic algorithm that works in the general framework of Moser and Tardos. We also give a parallel NC algorithm for the same setting, improving upon an algorithm of Alon [Random Structures Algorithms, 2 (1991), pp. 367--378].

The cutting plane approach to optimal matchings has been discussed by several authors over the past decades, and its rate of convergence has been an open question. We prove that the cutting plane approach using Edmonds' blossom inequalities converges in polynomial time for the minimum-cost perfect matching problem. Our main insight is an LP-based method to select cutting planes. This cut selection procedure leads to a sequence of intermediate linear programs with a linear number of constraints whose optima are half-integral and supported by a disjoint union of odd cycles and edges. This structural property of the optima is instrumental in finding violated blossom inequalities (cuts) in linear time. Moreover, the number of cycles in the support of the half-integral optima acts as a potential function to show efficient convergence to an integral solution.

A hitting set for a collection of sets is a set that has a nonempty intersection with each set in the collection; the hitting set problem is to find a hitting set of minimum cardinality. Motivated by instances of the hitting set problem where the number of sets to be hit is large, we introduce the notion of implicit hitting set problems. In an implicit hitting set problem the collection of sets to be hit is typically too large to list explicitly; instead, an oracle is provided which, given a set H, either determines that H is a hitting set or returns a set that H does not hit. We show a number of examples of classic implicit hitting set problems, and give a generic algorithm for solving such problems optimally. The main contribution of this paper is to show that this framework is valuable in developing approximation algorithms. We illustrate this methodology by presenting a simple on-line algorithm for the minimum feedback vertex set problem on random graphs. In particular our algorithm gives a feedback vertex set of size n–(1/p) log np(1 − o(1)) with probability at least 3/4 for the random graph Gn,p (the smallest feedback vertex set is of size n − (2/p) log np(1 + o(1))). We also consider a planted model for the feedback vertex set in directed random graphs. Here we show that a hitting set for a polynomial-sized subset of cycles is a hitting set for the planted random graph and this allows us to exactly recover the planted feedback vertex set.

Star-shaped bodies are an important nonconvex generalization of convex bodies (e.g., linear programming with violations). Here we present an efficient algorithm for sampling a given star-shaped body. The complexity of the algorithm grows polynomially in the dimension and inverse polynomially in the fraction of the volume taken up by the kernel of the star-shaped body. The analysis is based on a new isoperimetric inequality. Our main technical contribution is a tool for proving such inequalities when the domain is not convex. As a consequence, we obtain a polynomial algorithm for computing the volume of such a set as well. In contrast, linear optimization over star-shaped sets is NP-hard.

The Lovász Local Lemma [5] (LLL) is a powerful result in probability theory that states that the probability that none of a set of bad events happens is nonzero if the probability of each event is small compared to the number of events that depend on it. It is often used in combination with the probabilistic method for non-constructive existence proofs. A prominent application is to k-CNF formulas, where LLL implies that, if every clause in the formula shares variables with at most d ≤ 2k/e other clauses then such a formula has a satisfying assignment. Recently, a randomized algorithm to efficiently construct a satisfying assignment was given by Moser [13]. Subsequently Moser and Tardos [14] gave a randomized algorithm to construct the structures guaranteed by the LLL in a very general algorithmic framework. We address the main problem left open by Moser and Tardos of derandomizing these algorithms efficiently. Specifically, for a k-CNF formula with m clauses and d ≤ 2k/(1+ε)/e for some ε ∊ (0, 1), we give an algorithm that finds a satisfying assignment in time . This improves upon the deterministic algorithms of Moser and of Moser-Tardos with running times mΩ(k2) and mΩ(k · 1.ε) which are superpolynomial for k = ω(1) and upon other previous algorithms which work only for d ≤ 2k/16/e. Our algorithm works efficiently for the asymmetric version of LLL under the algorithmic framework of Moser and Tardos [14] and is also parallelizable, i.e., has polylogarithmic running time using polynomially many processors.