Leader-less ants make super efficient networks

Leader-less ants make super efficient networks
Argentine ants connect three nests in an empty arena via the shortest possible network. The ants have created a difficult Steiner network by adding an extra hub in the centre of the triangle, thus creating the network of absolute shortest trail length. The computer generated Steiner network for three nests is depicted in the square inset. Credit: Tanya Latty
(PhysOrg.com) -- Ants are able to connect multiple sites in the shortest possible way, and in doing so, create efficient transport networks, according to a University of Sydney study published in the Journal of the Royal Society Interface.

The research also revealed the process by which the ants solve network design problems without the help of a leader.

Dr. Tanya Latty, principal author from the School of Biological Sciences said the ants make as many trails as possible, then prune them back to the most efficient configuration.

"The findings sheds light on how other 'simple' natural systems without leaders or even brains - such as fungi, slime molds and mammalian vascular systems - are able to form efficient networks, and can help humans design artificial networks in situations lacking central control," she said.

"Engineers and urban planners face the task of designing efficient and cost effective networks. Building longer roads or tracks requires more resources and is therefore more costly, so a challenge for engineers is to design transportation networks that minimise resource use while still maintaining connectivity between sites such as cities or stations.

"Argentine ants face the same dilemma as transport engineers. This species of ant is a highly invasive pest in many countries because it can form super colonies that consist of thousands of nests connected by a network of pheromone trails. Because longer trails require more pheromone to build and maintain, the ants would benefit greatly from building efficient networks with the shortest possible trail length."

University of Sydney Associate Professor Madeleine Beekman, and co-author, who worked with the team of researchers from the University of Leipzig, Hiroshima University, Hokkaido Future University and University of Uppsala said the team posed the following question.

"Are ants able to solve this network efficiency problem given they have no leader, planner or anybody with global knowledge of the whole environment? And if so, how do they do it?" she said.

To see whether Argentine ants (Linepithima humile) could create inter-nest networks that minimized trail length, the researchers gave ants the task of connecting three or four nests together. The nests were placed in an empty arena and the ants were allowed to connect them in any way possible. After two hours, the resulting ant networks were photographed and compared to computer-generated images of networks that used the shortest trail length.

For each treatment, there were only two 'efficient' solutions that resulted in the shortest networks: the Minimal Spanning Tree, which connected the nests in the shortest way without adding an extra hub; and the Steiner Network, which used added hubs to connect the nests in the absolute shortest trail length. The Minimal Spanning tree is the simplest solution, but the Steiner network, with its added hubs, represents the most difficult but ultimately the most efficient route.

After two hours, the ants had created networks that closely resembled both efficient solutions. They frequently created difficult Steiner Networks, adding a central hub in the three-nest treatment and two central hubs in the four-nest treatment.

Dr. Latty said: "It's interesting that the ants were able to create, on a blank slate, the mathematically shortest network between multiple points. This network design allows Argentine ants to distribute brood, workers and food between nests with extreme efficiency, and might help explain why they viciously out-compete other ant species outside their native home of South America.

"But just because the ants are as good as computers at solving network design problems, doesn't mean that they use complex processes. We videotaped the ants forming their networks to work out exactly how they create this efficient solution, and found that networks formed through trial and error. Initially, ants built inefficient networks that had many redundant trails and dead ends but over the next few hours, redundant trails gradually disappeared until the reached its final, efficient configuration.

"So even though the ant colony seems to behave intelligently, the ants are simple individuals following even simpler rules. Understanding how simple organisms like build efficient networks can inform the design of human transportation networks."

Explore further

Next generation of algorithms inspired by problem-solving ants

Provided by University of Sydney
Citation: Leader-less ants make super efficient networks (2011, February 11) retrieved 25 June 2019 from https://phys.org/news/2011-02-leader-less-ants-super-efficient-networks.html
This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no part may be reproduced without the written permission. The content is provided for information purposes only.

Feedback to editors

User comments

Feb 11, 2011
I would say that insights gained from this would not just extend to transportation networks, biological neural networks also appear to follow similar processes: many random paths at first, slowly pruned down to reduce efficiency?

Feb 11, 2011
Computer routing networks that need to route packets via new path after network split could also benefit greatly from this research.

Feb 11, 2011
I don't get it. The ants are having to travel farther in this picture than if they had simply made straight line paths between the nodes. How is this advantageous in any way?

Feb 11, 2011
In other words, the optimization here is on total path length, not path length or transit time between two nests.
Right, but I think you're missing the part about the cost of pheromones. Apparently those chemicals are more energetically expensive than the sugars ants burn while moving. That would justify optimization of total path length, rather than time/distance of individual transits.

Feb 11, 2011
in any respect the most efficient route is a straight line -- A:B or A:C
Again, perhaps not in *every* respect. For instance, perhaps ants avoid laying down additional (expensive) pheromones when the pheromone trail is already strong. Let's further postulate that pheromone trails decay over time, and must be continually reinforced.

Under those assumptions, for the colony as a whole, it would cost more to maintain a Minimal Spanning Tree as opposed to a Steiner net.

Feb 11, 2011
Not arguing, so much as clarifying. The optimality of a network depends on the criteria of efficiency. If efficiency is maximized by reducing total sum of all path lengths (and thus minimizing overall quantity of pheromones required), then the most efficient routes may not be straight lines.

Feb 11, 2011
Graph Theory FTW ;) Have they uncovered the basic rules followed by each ant? I can't access the article from here.

Feb 11, 2011
American ants just outsource the path building to Chinese ants.

Feb 12, 2011
We do the same sometimes. I read about a place once that moved sidewalks, might have been a campus. When they found footpaths that took shortcuts they put sidewalks where the footpaths were. Smart and efficient and I'm sure it isn't done elsewhere, we aren't that smart as a species.

There is a little island in the road on the way to work that I have driven by for over 20 years. It has curbing that I've seen repaired over and over because one small corner gets driven over constantly. It always brought to mind the story above. Yet in all these years I've watched it they never move the curbing that one foot so that people don't drive over it, they just keep repairing it. stoopid monkeys

Feb 12, 2011
"When they found footpaths that took shortcuts they put sidewalks where the footpaths were"

That's great..! Can you remember where this was? It would make sense to build more foodpaths this way.

Feb 13, 2011
This is an emergent system; I wouldn't characterize it as 'simple'. Us humans naturally think in terms of process, but emergent systems are borne of rules, not process. Ants are truly amazing in that a single ant is simple and incapable, but a colony is capable.

In further study, I would like to know the number of ants it takes for this system to form.

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more