Computer Science > Data Structures and Algorithms
[Submitted on 29 Mar 2020 (v1), revised 9 Apr 2020 (this version, v2), latest version 21 Jul 2021 (v3)]
Title:The Search Efficiency of Intermittent Lévy walks Optimally Scales with Target Size
View PDFAbstract:Lévy walks are random walks processes whose step length follows a long-tailed power law distribution. Due to their abundance as movement patterns of biological organisms, significant theoretical efforts have been devoted to identify the foraging circumstances that would make such patterns advantageous [Viswanathan et al. Nature, 1999]. Despite numerous attempts, there is currently no conclusive analytical evidence that Lévy flights are preferable strategies in higher dimensions than one. Here we show that the optimality of inverse-square Lévy walks in two-dimensions becomes striking when targets are sparse and unpredictable in size, and when detection is weak. Specifically, we prove that under the intermittent model, in which detection is not possible while moving ballistically [Bénichou et al. Reviews of Modern Physics, 2011], this strategy optimally finds sparse targets of any size and shape. That is, in a square torus of area~$n$, and assuming that the detection radius is normalized to $1$, the strategy finds any connected set of diameter $D$ in $\tilde{O}(n/D)$ expected time, whereas $\Omega(n/D)$ is an unconditional lower bound on the expected time, that holds even when assuming that the shape and size of the target are known. Furthermore, this particular Lévy process stands in stark contrast to many other basic intermittent processes, including all other Lévy walks, whose competitive ratio is shown to be polynomial in $n$, for wide ranges of target scales. Our results thus provide strong theoretical support for the optimality and robustness of intermittent Lévy walks under general conditions.
Submission history
From: Amos Korman [view email][v1] Sun, 29 Mar 2020 14:45:02 UTC (116 KB)
[v2] Thu, 9 Apr 2020 10:14:36 UTC (117 KB)
[v3] Wed, 21 Jul 2021 22:30:27 UTC (244 KB)
Current browse context:
cs.DS
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.