### References & Citations

# Computer Science > Data Structures and Algorithms

# Title: A Dynamic Data Structure for Temporal Reachability with Unsorted Contact Insertions

(Submitted on 8 Feb 2021 (v1), last revised 30 Mar 2021 (this version, v2))

Abstract: Temporal graphs represent interactions between entities over the time. These interactions may be direct (a contact between two nodes at some time instant), or indirect, through sequences of contacts called temporal paths (journeys). Deciding whether an entity can reach another through a journey is useful for various applications in communication networks and epidemiology, among other fields. In this paper, we present a data structure which maintains temporal reachability information under the addition of new contacts (i.e., triplets $(u,v,t)$ indicating that node $u$ and node $v$ interacted at time $t$). In contrast to previous works, the contacts can be inserted in arbitrary order -- in particular, non-chronologically -- which corresponds to systems where the information is collected a posteriori (e.g. when trying to reconstruct contamination chains among people). The main component of our data structure is a generalization of transitive closure called timed transitive closure (TTC), which allows us to maintain reachability information relative to all nested time intervals, without storing all these intervals, nor the journeys themselves. TTCs are of independent interest and we study a number of their general properties. Let $n$ be the number of nodes and $\tau$ be the number of timestamps in the lifetime of the temporal graph. Our data structure answers reachability queries regarding the existence of a journey from a given node to another within given time interval in time $O(\log\tau)$; it has an amortized insertion time of $O(n^2\log\tau)$; and it can reconstruct a valid journey that witnesses reachability in time $O(k\log\tau)$, where $k<n$ is the maximum number of edges of this journey. Finally, the space complexity of our reachability data structure is $O(n^2\tau)$, which remains within the worst-case size of the temporal graph itself.

## Submission history

From: Luiz Fernando Afra Brito [view email]**[v1]**Mon, 8 Feb 2021 13:41:36 GMT (35kb)

**[v2]**Tue, 30 Mar 2021 19:18:27 GMT (36kb)

Link back to: arXiv, form interface, contact.