Impact of network topology changes on information source localization

TL;DR

This study evaluates how different network topology modifications affect the performance of source localization algorithms, revealing that GMLA is resilient to added links, while Pearson excels when links are hidden or reattached.

Contribution

The paper compares the robustness of three localization algorithms under various network topology alterations, highlighting the strengths and weaknesses of each method.

Findings

GMLA is highly resilient to added superfluous edges.

Pearson algorithm performs better when links are hidden or reattached.

GMLA is more accurate and significantly faster than LPTVA.

Abstract

Well-established methods of locating the source of information in a complex network are usually derived with the assumption of complete and exact knowledge of network topology. We study the performance of three such algorithms (LPTVA, GMLA and Pearson correlation algorithm) in scenarios that do not fulfill this assumption by modifying the network prior to localization. This is done by adding superfluous new links, hiding existing ones, or reattaching links in accordance with the network's structural Hamiltonian. We find that GMLA is highly resilient to the addition of superfluous edges, as its precision falls by more than statistical uncertainty only when the number of links is approximately doubled. On the other hand, if the edge set is underestimated or reattachment has taken place, the performance of GMLA drops significantly. In such a scenario the Pearson algorithm is preferable, retaining most of its performance when other simulation parameters favor localization (high density of observers, highly deterministic propagation). It is also generally more accurate than LPTVA, as well as orders of magnitude faster. The aforementioned differences between localization algorithms can be intuitively explained, although a need for further theoretical research is noted.