The Traveling Salesman Problem (TSP) in the $d$-dimensional Euclidean space is among the oldest and most famous NP-hard optimization problems. In breakthrough works, Arora [J. ACM 1998] and Mitchell [SICOMP 1999] gave the first polynomial time approximation schemes. To improve the running time, Rao and Smith [STOC 1998] gave a randomized $(1/\varepsilon)^{O(1/\varepsilon^{d-1})}\cdot n\log n$ time approximation scheme. Bartal and Gottlieb [FOCS 2013] gave a randomized approximation scheme in $2^{(1/\varepsilon)^{O(d)}} n$ time, which is linear in $n$. Recently, Kisfaludi-Bak, Nederlof, and Węgrzycki [FOCS 2021] gave a randomized approximation scheme in $2^{O(1/\varepsilon^{d-1})} n \log n$ time, achieving a Gap-ETH tight dependence on $\varepsilon$. It is raised as a challenging open question by Kisfaludi-Bak, Nederlof, and Węgrzycki [FOCS 2021] whether a running time of $2^{O(1/\varepsilon^{d-1})}n$ is achievable. We answer their question positively by giving a randomized $2^{O(1/\varepsilon^{d-1})} n$ time approximation scheme for Euclidean TSP.

Oriented spanners were presented at ESA'23 as an extension of the well-researched geometric spanners: Given a set P of points in a metric space and an oriented graph G, the oriented dilation of two points p,q ∈ P is the length of the shortest closed walk in G containing p and q divided by the minimum perimeter triangle of p and q. G is called a t-spanner, if the maximum dilation over all pairs of points in P is at most t. This paper presents the first constructions of sparse oriented spanners for metric spaces beyond the Euclidean space. Given an orientation of the complete graph (i.e. a tournament) with dilation t on n points that satisfies an additional short-cycle property, we show how to extract a (t+ε)-spanner with 𝒪(k) edges in 𝒪(kn²+T(n)) time, for any metric space admitting a well-separated pair decomposition with k pairs computable in T(n) time. We supplement this with an improved construction of tournaments for metric point sets, obtaining dilation 5/3. This improves the previous bound of 2 and approaches the lower bound of 1.5. Combined, for n points in a metric space with constant doubling dimension d, this yields a (5/3 + ε)-spanner with (1/ε)^{𝒪(d)}n edges computable in (1/ε)^𝒪(d) n³ time using 𝒪(n²) space. This improves the dilation over the (2+ε)-spanner for Euclidean point sets presented at SoCG’25 while applying to more general metric spaces. Moreover, we generalize the known (2+ε)-spanner to doubling spaces. In particular, an oriented (2+ε)-spanner with 𝒪(ε^{-d} n) edges can be constructed in (1/ε)^𝒪(d) n log n time using 𝒪(ε^{-d} n) space. Since the oriented dilation can be dominated by one pair of points, we also consider the oriented average dilation, which is the sum over the oriented dilation of all pairs of points divided by the number of pairs. While oriented (1+ε)-spanners do not exist for every point set, we present an algorithm that computes a spanner with average dilation 1+ε for point sets in a metric space of constant doubling dimension d: More concretely, our algorithm computes an oriented spanner with average dilation at most 1 + 𝒪(1/s) + s^𝒪(d)/n with s^𝒪(d) n edges in s^𝒪(d) n log n time using s^𝒪(d) n space, where s is any sufficiently large number that may depend on n.

The shifting technique of Hochbaum and Maass [J.ACM'85] produces PTASes with the fastest known running times n^O(1/ε^{d-1}) for several d dimensional geometric problems. However, it is only known, due to Marx [FOCS'07], that these algorithms are indeed optimal for dimension d = 2. We show that these running times are optimal under Gap-ETH for every constant dimension. More precisely, we develop a framework that enables us to prove the conditional optimality of the shifting algorithms for several problems on unit ball graphs, such as maximum independent set, maximum induced forest, and others, as well as for the problem of piercing unit balls. Our framework is built using the cube wiring theorem of De Berg et al. [SICOMP'20] and the reduction steps of Marx and Sidiropoulos [SoCG'14] to create a convenient maximization version of geometric CSP that can be used as a basis for reductions.

A dominating set of a graph G is connected if it induces a connected graph in G. For planar triangulations, it has been known since 1990 that every n-vertex triangulation admits a connected dominating set of size at most n/2 - 1, and no improvement to this bound was known for over three decades. We break this longstanding barrier by showing that every n-vertex triangulation has a connected dominating set of size at most 10n/21. Equivalently, every triangulation admits a spanning tree with at least 11n/21 leaves. Moreover, we present an algorithm that computes such a set in optimal linear time. Our result narrows the gap to the best known lower bound and has graph drawing applications, establishing a bound for one-bend free sets and improving the known bound for simultaneous planar embeddings.

We introduce the contiguous art gallery problem which is to guard the boundary of a simple polygon with a minimum number of guards such that each guard covers exactly one contiguous portion of the boundary. Art gallery problems are often NP-hard. In particular, it is NP-hard to minimize the number of guards to see the boundary of a simple polygon, without the contiguity constraint. This paper is a merge of three concurrent works [Ahmad Biniaz et al., 2024; Magnus Christian Ring Merrild et al., 2024; Eliot W. Robson et al., 2024] each showing that (surprisingly) the contiguous art gallery problem is solvable in polynomial time. The common idea of all three approaches is developing a greedy function that maps a point on the boundary to the furthest point on the boundary so that the contiguous interval along the boundary between them could be guarded by one guard. Repeatedly applying this function immediately leads to an OPT+1 approximation. By studying this greedy algorithm, we present three different approaches that achieve an optimal solution. The first and second approach apply this greedy algorithm from different points on the boundary that could be found in advance or on the fly while traversing along the boundary (respectively). The third approach represents this function as a piecewise linear rational function, which can be reduced to an abstract arc cover problem involving infinite families of arcs. We identify other problems that can be represented by similar functions, and solve them via the third approach. From the combinatorial point of view, we show that any n-vertex polygon can be guarded by at most ⌊(n-2)/2⌋ guards. This bound is tight because there are polygons that require this many guards.

Given a point set P in a metric space and a real number t ≥ 1, an oriented t-spanner is an oriented graph G = (P, E), where for every pair of distinct points p and q in P, the shortest oriented closed walk in G that contains p and q is at most a factor t longer than the perimeter of the smallest triangle in P containing p and q. The oriented dilation of a graph G is the minimum t for which G is an oriented t-spanner. For arbitrary point sets of size n in ℝ^d, where d ≥ 2 is a constant, the only known oriented spanner construction is an oriented 2-spanner with binom(n,2) edges. Moreover, there exists a set P of four points in the plane, for which the oriented dilation is larger than 1.46, for any oriented graph on P. We present the first algorithm that computes, in Euclidean space, a sparse oriented spanner whose oriented dilation is bounded by a constant. More specifically, for any set of n points in ℝ^d, where d is a constant, we construct an oriented (2+ε)-spanner with 𝒪(n) edges in 𝒪(n log n) time and 𝒪(n) space. Our construction uses the well-separated pair decomposition and an algorithm that computes a (1+ε)-approximation of the minimum-perimeter triangle in P containing two given query points in 𝒪(log n) time. While our algorithm is based on first computing a suitable undirected graph and then orienting it, we show that, in general, computing the orientation of an undirected graph that minimises its oriented dilation is NP-hard, even for point sets in the Euclidean plane. We further prove that even if the oriented graph is already given, computing its oriented dilation is APSP-hard for points in a general metric space. We complement this result with an algorithm that approximates the oriented dilation of a given graph in subcubic time for point sets in ℝ^d, where d is a constant.

Let S be a set of n points in ℝ^d, where d ≥ 2 is a constant, and let H₁,H₂,…,H_{m+1} be a sequence of vertical hyperplanes that are sorted by their first coordinates, such that exactly n/m points of S are between any two successive hyperplanes. Let |A(S,m)| be the number of different closest pairs in the {(m+1) choose 2} vertical slabs that are bounded by H_i and H_j, over all 1 ≤ i < j ≤ m+1. We prove tight bounds for the largest possible value of |A(S,m)|, over all point sets of size n, and for all values of 1 ≤ m ≤ n. As a result of these bounds, we obtain, for any constant ε > 0, a data structure of size O(n), such that for any vertical query slab Q, the closest pair in the set Q ∩ S can be reported in O(n^{1/2+ε}) time. Prior to this work, no linear space data structure with sublinear query time was known.