In a linear combinatorial optimization problem, we are given a family F ⊆ 2U of feasible subsets of a ground set U of n elements, and our goal is to find S* = argminS ∈ F ⟨ w,1S ⟩. Traditionally, we are either given the weight vector up-front, or else we are given a value oracle which allows us to evaluate w(S) := ⟨ w, 1S ⟩ for any S ∈ F. We consider the weaker and more robust comparison oracle, which for any two feasible sets S, T ∈ F, reveals only if w(S) is less than/equal to/greater than w(T). We ask: When can we find the optimal feasible set S* = argminS ∈ F w(S) using a small number of comparison queries? If so, when can we do this efficiently?

We study the classic Max-Cut problem under multiple cardinality constraints, which we refer to as the Constrained Max-Cut problem. Given a graph G = (V, E), a partition of the vertices into c disjoint parts V₁, …, V_c, and cardinality parameters k₁, …, k_c, the goal is to select a set S ⊆ V such that |S ∩ V_i| = k_i for each i ∈ [c], maximizing the total weight of edges crossing S (i.e., edges with exactly one endpoint in S). By designing an approximate kernel for Constrained Max-Cut and building on the correlation rounding technique of Raghavendra and Tan (2012), we present a (0.858 - ε)-approximation algorithm for the problem when c = O(1). The algorithm runs in time O(min{k/ε, n}^poly(c/ε) + poly(n)), where k = ∑_{i∈[c]} k_i and n = |V|. This improves upon the (1/2 + ε₀)-approximation of Feige and Langberg (2001) for Max-Cut_k (the special case when c = 1, k₁ = k), and generalizes the (0.858 - ε)-approximation of Raghavendra and Tan (2012), which only applies when min{k,n-k} = Ω(n) and does not handle multiple constraints. We also establish that, for general values of c, it is NP-hard to determine whether a feasible solution exists that cuts all edges. Finally, we present a 1/2-approximation algorithm for Max-Cut under an arbitrary matroid constraint.

In this paper we study constrained subspace approximation problem. Given a set of n points {a₁,…,a_n} in ℝ^d, the goal of the subspace approximation problem is to find a k dimensional subspace that best approximates the input points. More precisely, for a given p ≥ 1, we aim to minimize the pth power of the 𝓁_p norm of the error vector (‖a₁-Pa₁‖,…,‖a_n-Pa_n‖), where P denotes the projection matrix onto the subspace and the norms are Euclidean. In constrained subspace approximation (CSA), we additionally have constraints on the projection matrix P. In its most general form, we require P to belong to a given subset 𝒮 that is described explicitly or implicitly. We introduce a general framework for constrained subspace approximation. Our approach, that we term coreset-guess-solve, yields either (1+ε)-multiplicative or ε-additive approximations for a variety of constraints. We show that it provides new algorithms for partition-constrained subspace approximation with applications to fair subspace approximation, k-means clustering, and projected non-negative matrix factorization, among others. Specifically, while we reconstruct the best known bounds for k-means clustering in Euclidean spaces, we improve the known results for the remainder of the problems.

In an instance of the weighted Nash Social Welfare problem, we are given a set of m indivisible items, G, and n agents, A, where each agent i ∈ A has a valuation vij ≥ 0 for each item j ∈ G. In addition, every agent i has a non-negative weight wi such that the weights collectively sum up to 1. The goal is to find an assignment σ : G → A that maximizes . When all the weights equal to , the problem reduces to the classical Nash Social Welfare problem, which has recently received much attention. In this work, we present a -approximation algorithm for the weighted Nash Social Welfare problem, where denotes the KL-divergence between the distribution w and the uniform distribution on [n].

Given a set of points in d-dimensional space, an explainable clustering is one where the clusters are specified by a tree of axis-aligned threshold cuts. Dasgupta et al. (ICML 2020) posed the question of the price of explainability: the worst-case ratio between the cost of the best explainable clusterings to that of the best clusterings.We show that the price of explainability for k medians is at most $1+H_{k-1}$; in fact, we show that the popular Random Thresholds algorithm has exactly this price of explainability, matching the known lower bound constructions. We complement our tight analysis of this particular algorithm by constructing instances where the price of explainability (using any algorithm) is at least $(1-o(1)) \ln k$, showing that our result is best possible, up to lower-order terms. We also improve the price of explainability for the k-means problem to $O(k \ln \ln k)$ from the previous $O(k \ln k)$, considerably closing the gap to the lower bounds of $\Omega(k)$. Finally, we study the algorithmic question of finding the best explainable clustering: We show that explainable k medians and k-means cannot be approximated better than $O(\ln k)$, under standard complexity-theoretic conjectures. This essentially settles the approximability of explainable k-medians and leaves open the intriguing possibility to get significantly better approximation algorithms for k-means than its price of explainability.

Determinant maximization problem gives a general framework that models problems arising in as diverse fields as statistics [1], convex geometry [2], fair allocations [3], combinatorics [4], spectral graph theory [5], network design, and random processes [6]. In an instance of a determinant maximization problem, we are given a collection of vectors $U=\{v_{1},\cdots,\ v_{n}\}\subset \mathbb{R}^{d}$, and a goal is to pick a subset $S\subseteq U$ of given vectors to maximize the determinant of the matrix $\displaystyle \sum_{i\in S}v_{i}v_{i}^{\text{T}}$. Often, the set S of picked vectors must satisfy additional combinatorial constraints such as cardinality constraint $(|S|\leq k)$ or matroid constraint $(S$ is a basis of a matroid defined on the vectors). In this paper, we give a polynomial-time deterministic algorithm that returns a $r^{O(r)}$-approximation for any matroid of rank $r \leq d$. This improves previous results that give $e^{O(r^{2})}$-approximation algorithms relying on $e^{O(r)}$-approximate estimation algorithms [4], [7] –[9] for any r$\leq$d. All previous results use convex relaxations and their relationship to stable polynomials and strongly $\log$-concave polynomials or non-convex relaxations for the problem [10]. In contrast, our algorithm builds on combinatorial algorithms for matroid intersection, which iteratively improve any solution by finding an alternating negative cycle in the exchange graph defined by the matroids. While the $\det(.)$ function is not linear, we show that taking appropriate linear approximations at each iteration suffice to give the improved approximation algorithm.

We study the Maximum Independent Set of Rectangles (MISR) problem, where we are given a set of axis-parallel rectangles in the plane and the goal is to select a subset of non-overlapping rectangles of maximum cardinality. In a recent breakthrough, Mitchell [46] obtained the first constant-factor approximation algorithm for MISR. His algorithm achieves an approximation ratio of 10 and it is based on a dynamic program that intuitively recursively partitions the input plane into special polygons called corner-clipped rectangles (CCRs), without intersecting certain special horizontal line segments called fences. In this paper, we present a 3-approximation algorithm for MISR which is also based on a recursive partitioning scheme. First, we use a partition into a class of axis-parallel polygons with constant complexity each that are more general than CCRs. This allows us to provide an arguably simpler analysis and at the same time already improves the approximation ratio to 6. Then, using a more elaborate charging scheme and a recursive partitioning into general axis-parallel polygons with constant complexity, we improve our approximation ratio to 3. In particular, we construct a recursive partitioning based on more general fences which can be sequences of up to O(1) line segments each. This partitioning routine and our other new ideas may be useful for future work towards a PTAS for MISR.

Guillotine separability of rectangles has recently gained prominence in combinatorial optimization, computational geometry, and combinatorics. Consider a given large stock unit (say glass or wood) and we need to cut out a set of required rectangles from it. Many cutting technologies allow only end-to-end cuts called guillotine cuts. Guillotine cuts occur in stages. Each stage consists of either only vertical cuts or only horizontal cuts. In k-stage packing, the number of cuts to obtain each rectangle from the initial packing is at most k (plus an additional trimming step to separate the rectangle itself from a waste area). Pach and Tardos [Pach and Tardos, 2000] studied the following question: Given a set of n axis-parallel rectangles (in the weighted case, each rectangle has an associated weight), cut out as many rectangles (resp. weight) as possible using a sequence of guillotine cuts. They provide a guillotine cutting sequence that recovers 1/(2 log n)-fraction of rectangles (resp. weights). Abed et al. [Fidaa Abed et al., 2015] claimed that a guillotine cutting sequence can recover a constant fraction for axis-parallel squares. They also conjectured that for any set of rectangles, there exists a sequence of axis-parallel guillotine cuts that recovers a constant fraction of rectangles. This conjecture, if true, would yield a combinatorial O(1)-approximation for Maximum Independent Set of Rectangles (MISR), a long-standing open problem. We show the conjecture is not true, if we only allow o(log log n) stages (resp. o(log n/log log n)-stages for the weighted case). On the positive side, we show a simple O(n log n)-time 2-stage cut sequence that recovers 1/(1+log n)-fraction of rectangles. We improve the extraction of squares by showing that 1/40-fraction (resp. 1/160 in the weighted case) of squares can be recovered using guillotine cuts. We also show O(1)-fraction of rectangles, even in the weighted case, can be recovered for many special cases of rectangles, e.g. fat (bounded width/height), δ-large (large in one of the dimensions), etc. We show that this implies O(1)-factor approximation for Maximum Weighted Independent Set of Rectangles, the weighted version of MISR, for these classes of rectangles.