Paper Info
Title
Decentralized Congestion Control in Random Ant Interaction Networks.
Abstract
Interaction networks formed by foraging ants are among the most studied self-organizing multi-agent systems in nature that have inspired many practical applications. However, the vast majority of prior investigations assume pheromone trails or stigmergic strategies used by the ants to create foraging behaviors. We first review an ant network model where the direction and speed of each ant's correlated random walk are influenced by direct and minimalist interactions, such as antennal contact. We incorporate basic ant memory with nest and food compasses, and adopt a discrete time, non-deterministic forager recruitment strategy to regulate the foraging population. The paper's main focus is on decentralized congestion control and avoidance schemes that are activated with a quorum sensing mechanism. The model relies on individual ants' ability to estimate a perceived avoidance sector from recent interactions. Through simulation experiments it is shown that a randomized congestion avoidance scheme improves performance over alternative static schemes.
Year
DOI
Venue
2017
10.1007/978-3-319-61824-1_29
ADVANCES IN SWARM INTELLIGENCE, ICSI 2017, PT I
Field
DocType
Volume
ANT,Population,Random walk,Computer science,Network congestion,Discrete time and continuous time,Nest,Foraging,Distributed computing
Conference
10385
ISSN
Citations 
PageRank 
0302-9743
0
0.34
References 
Authors
7
3
Name
Order
Citations
PageRank
Andreas Kasprzok100.34
Beshah Ayalew25612.79
Chad Lau310.68