CS Paper Radar

The Tight Competitive Ratio of Greedy for Online Steiner Forest

Iman Gholami, Ilya Hajiaghayi, MohammadTaghi Hajiaghayi, Peyman Jabbarzade

cs.DS ·submitted 2026-10-29 17:43 UTC · v1
theory

Abstract

The online Steiner Forest problem is a fundamental network design problem in which terminal pairs arrive one at a time and an algorithm must irrevocably purchase edges connecting each pair. The natural greedy algorithm connects each terminal pair by a shortest path after making all previously purchased edges free. Awerbuch, Azar, and Bartal (SODA 1996) proved that this algorithm is $O(log^2 k)$-competitive and conjectured an $O(log k)$ bound, where $k$ is the number of terminal pairs. Bamas, Drygala, and Maggiori (SODA 2022) obtained an $O(\log \NumPairs\log\log \NumPairs)$ bound for greedy on a restricted class of inputs, leaving the general case open. In this paper, we resolve the competitive ratio of the greedy algorithm for online Steiner Forest, proving that it is tightly $Θ(log(k) log log(k))$. Specifically, we establish an $O(log(k) log log(k))$ upper bound for arbitrary instances and construct a matching $Ω(log(k) log log(k))$ lower bound. Our results refute the $O(log k)$ conjecture of Awerbuch, Azar, and Bartal and establish a tight $Θ(log log k)$ separation between the greedy algorithm and the best possible online algorithm for Steiner Forest.

Closest papers

Nearest neighbours by meaning across the whole index.

Hardness of Online Directed Steiner Network

cs.DS ·2026-09-18 ·Gary Hoppenworth, Yaowei Long, Sepideh Mahabadi, Jakub Tarnawski
similarity 0.73
Abstract

In the Directed Steiner Network (DSN) problem we are given a directed graph and a set of demands $(s_i,t_i)$, and asked to find a cheap subgraph connecting each terminal pair. In its online version, the demands arrive online and must be served by buying edges irrevocably. DSN is a fundamental hard problem in network design, heavily studied in both the offline and the online setting. Offline, it has a superpolylogarithmic hardness of approximation. However, offline hardness says nothing about online algorithms, which are computationally unrestricted. It has been an open question whether uncertainty itself (needing to commit to a solution without knowing future demands) rules out polylogarithmic-competitive online algorithms. In this work, we show the first such unconditional, information-theoretic hardness. Namely, we give an $\exp\!\bigl(Ω(\sqrt{\log n})\bigr)$ bound on the competitive ratio, which holds even for randomized algorithms against an oblivious adversary, and on unit-cost DAGs. Our proof uses a novel connection between online network design and algebraic coding theory. We encode requests using a hidden low-degree polynomial, whose past evaluations reveal nothing about future ones. We then use list-recovery bounds to show that an algorithm cannot make cheaply reusable decisions without knowing those future evaluations.

0.18

Optimal Online Weighted Connectivity Augmentation

cs.DS ·2026-10-30 ·Mohit Garg, Felix Hommelsheim, Alexander Lindermayr
theory similarity 0.69
Abstract · 32 pages, 12 figures

In the online Weighted Connectivity Augmentation Problem (WCAP), we are given a $k$-edge-connected graph on $n$ vertices and a set of additional edges, called links, with nonnegative costs. Pairs of vertices arrive one by one, and an online algorithm must immediately and irrevocably buy links so that each pair becomes $(k+1)$-edge-connected. The goal is to minimize the total cost of the purchased links. We obtain a deterministic polynomial-time $O(\log n)$-competitive algorithm for WCAP, improving the previous $O(\log^2 n)$ bound (SODA 2026). Our bound is asymptotically optimal, even for randomized algorithms in the random-order model. The general problem reduces to the case where the underlying graph is a cycle, for which link costs in $\{0,1\}$ already yield an $Ω(\log n)$ lower bound. Surprisingly, for unit costs on a cycle, we obtain a deterministic polynomial-time $4$-competitive algorithm, improving the previous $O(\log n)$ bound (SODA 2026), and prove that this ratio is optimal among deterministic algorithms. Thus, with unit costs, cycles are easier than trees, where the optimal competitive ratio is $Θ(\log n)$.

0.25

Online Survivable Network Design with Arbitrarily Small Connectivity Loss

cs.DS ·2026-11-12 ·MohammadTaghi Hajiaghayi, Saeed Seddighin
similarity 0.67
Abstract

In the online survivable network design problem, a graph with edge costs is known in advance, and requests arrive one at a time. Each request is a pair of vertices $(s,t)$ with a connectivity requirement $k$, and the algorithm must irrevocably buy edges so that $s$ and $t$ are joined by $k$ edge-disjoint paths. The goal is to keep the total cost close to that of an optimal offline solution. Gupta, Krishnaswamy, and Ravi~\cite{GKR12} obtain an $O(k\log^3 n)$-competitive algorithm, where $k$ is the maximum requirement, without relaxing the connectivity requirement. Dehghani, Ehsani, Hajiaghayi, Liaghat, and Seddighin~\cite{DEHLS18} obtain a polylogarithmic competitive ratio when the connectivity requirement is allowed to be relaxed by a factor of $2+\varepsilon$. We show that relaxing the connectivity requirement by a factor of only $1+\varepsilon$ suffices to obtain a polylogarithmic competitive ratio.

0.19

Sublogarithmic Online Graph Exploration

cs.DS cs.DM ·2026-11-10 ·Julia Baligacs
similarity 0.67
Abstract

We study the online graph exploration problem, in which an agent must visit every vertex of an initially unknown weighted graph on $n$ vertices and return to its starting vertex. In this model, a vertex's incident edges and their weights are revealed when it is first visited. Despite extensive research, the best known competitive guarantee for general graphs had remained $\mathcal{O}(\log n)$, a bound already established in 1977 for the greedy TSP heuristic Nearest Neighbor. Subsequent approaches, including hierarchical depth-first search (hDFS) and Blocking, achieve better guarantees on restricted graph classes but have logarithmic or worse competitive ratios on general graphs. This raises the question of whether the logarithmic barrier is inherent and Nearest Neighbor is asymptotically optimal. We answer this question in the negative by giving an $\mathcal{O}(\log n/\log \log n)$-competitive algorithm. Our algorithm is a slight modification of hDFS, and we establish its competitive guarantee with a surprisingly short proof.

0.22

A tight lower bound for malicious online bipartite matching with limited recourse budget

cs.DS cs.DM ·2026-08-23 ·Julia Baligacs, Bartłomiej Bosek, Paweł Putra, Marek Sokołowski, Anna Zych-Pawlewicz
theory similarity 0.67
Abstract

We study one-sided online bipartite matching with recourse. In this setting, one side of a bipartite graph is known in advance, while vertices on the other side arrive online together with their incident edges. After each arrival, the algorithm must maintain a maximum-cardinality matching while minimizing the total number of reallocations, also known as the recourse budget. Despite extensive work, the exact recourse complexity of the problem remains unsettled: the best lower bound is $Ω(n \log n)$, whereas the best upper bound is $\mathcal{O}(n \log^2 n)$, where $n$ denotes the number of online vertices. Tight upper bounds of $\mathcal{O}(n \log n)$ are known only for restricted graph classes, such as forests. The best known upper bounds are attained by a very simple and natural algorithm SAP, which after each arrival applies a shortest augmenting path, and it is conjectured to be optimal. All known upper bound analyses of this algorithm do not depend on the particular maximum matching maintained by the algorithm. Consequently, they also apply to a more difficult problem, which we call the malicious matching setting: after each arrival, the maintained matching is replaced by a worst-case maximum matching for the next step. This led to the conjecture that the malicious setting still admits an $\mathcal{O}(n \log n)$ recourse bound, in line with the conjectured optimal complexity of the original model. Our main result is an $Ω(n \log^2 n)$ lower bound for the malicious matching setting, thus disproving the conjecture. Together with the previous upper bound, this settles the asymptotic recourse complexity of the malicious variant of the problem. We complement our lower bound with an upper bound of $\mathcal{O}(n \log n)$ for expander graphs.

0.20

Online Vertex Connectivity Augmentation

cs.DS ·2026-11-11 ·Yaowei Long, Sepideh Mahabadi, Jakub Tarnawski
similarity 0.67
Abstract · 35 pages

Network design problems form a classic subfield of online algorithms, but almost nothing is known about their more challenging vertex-connectivity variants. We initiate the study of online vertex-connectivity network design for arbitrary demand pairs by considering the fundamental problem of connectivity augmentation. In an instance of \emph{Online Vertex Connectivity Augmentation} (OVCA), we are given a $k$-vertex-connected graph $F$ and a set of additional purchasable edges (with costs). Demand pairs $(s_1,t_1),(s_2,t_2),\dots$ then arrive online. Upon each arrival, the algorithm must irrevocably purchase edges so that $s_i$ and $t_i$ become $(k{+}1)$-vertex-connected in the union of $F$ with all edges purchased so far. Our main contribution is an $O((k+\log n)\log n)$-competitive online algorithm for OVCA. We also give an offline $O(k+\log n)$-approximation algorithm. We obtain our algorithms via a structural dichotomy on the minimum vertex cuts of $F$. We note that some dependence on $k$ in the approximation ratio is necessary -- i.e., the problem is hard to approximate within any $\mathrm{polylog}(n)$ factor independent of $k$ -- which shows a separation from the edge-connectivity setting. Finally, specializing to $k=1$, our methods yield the first polylogarithmic-competitive online algorithm for a vertex-connectivity survivable network design problem, namely an $\widetilde O(\log^3 n)$-competitive algorithm for online $\{1,2\}$-VC-SNDP (i.e., an SNDP variant where each demand pair comes with a vertex-connectivity requirement of $1$ or $2$).

0.18

A Sublogarithmic Competitive Algorithm for Online Graph Exploration

cs.DS ·2026-11-10 ·Max Klimm
similarity 0.66
Abstract

In online graph exploration, an agent must visit every vertex of an unknown weighted graph and return to its start, learning the incident edges at each vertex it visits. On $n$-vertex graphs, the best known competitive ratio has been $O(\log n)$, achieved by several algorithms, among them the hierarchical depth-first search of Megow, Mehlhorn, and Schweitzer. We give an algorithm with competitive ratio ${(2+o(1))}\ln n/\ln\ln n$ for known~$n$. Like hierarchical depth-first search, it groups the edges into geometric weight classes and searches depth-first between components of lighter edges. Unlike it, the agent tours each component along a minimum spanning tree for the original weights, so only crossings between components pay for rounding, and the base $b$ of the weight class can grow with~$n$. The leading constant $2$ is tight for our choice of base. Without knowing $n$, restarts with increasing size guesses give upper bounds with leading constant $4$ for a fixed schedule and $2$ in expectation for a randomly shifted schedule. If the weights lie in a known range $[1,2^k]$ with $k\ge4$, choosing $b=k/\log_2 k$ gives a competitive ratio of $O(k/\log k)$ without knowing $n$.

0.26

An $n^{0.3+\varepsilon}$-Approximation for Steiner $k$-Forest

cs.DS ·2026-10-07 ·Eden Chlamtáč
similarity 0.65
Abstract

We give an $n^{0.3+\varepsilon}$-approximation algorithm for the Steiner $k$-Forest problem, for any constant $\varepsilon>0$. As a function of $n$, this improves over the $O(\min\{\sqrt{n},\sqrt{k}\})$-approximation of Gupta et al. [ESA'07, TALG'10] which has stood for nearly two decades for the general case, as well as the later $n^{0.448}$-approximation of Dinitz et. al [APPROX-RANDOM'14, TALG'17] for the uniform weight case. On the other hand, we show that, due to lower bounds on the Densest $k$-Subgraph problem, the $O(\sqrt k)$-approximation for Steiner $k$-Forest likely cannot be improved. Specifically, we show that for any sufficiently small $\varepsilon>0$, an $O(k^{1/2-\varepsilon})$-approximation for Steiner $k$-Forest would surpass known degree-$n^{Ω(\varepsilon^2)}$ Sum-of-Squares integrality gaps for Densest k-Subgraph and refute the corresponding dense-versus-random conjecture.

0.20