Ant Behavior Understanding
Compact satellites, faster plane boarding and quicker downloads – the humble ant has inspired all three. Adam Hart reveals what these six-legged marvels can teach us.
To LOOK INSIDE an ant nest is to contemplate an alien civilization. The boiling mass of worker ants beneath an upturned stone is both strangely reminiscent of human society and strikingly different. There is an industry and organisation that fascinates us and a long line of myrmecophiles (or ant lovers) leads back all the way to King Solomon, who in fact advised to “go to the ant, consider her ways and be wise”. This was exactly the inspiration behind Planet Ant, a BBC Four programme showcasing what we know about the realm of ants, and what ants can teach us about the human world.
Like us, ants build structures, find food, defend their societies and manage waste, and – also like us – they must be well organized. For example, the leaf-cutting ants of Planet Ant have special waste disposal areas for storing hazardous waste and a team of ‘waste-disposal ants’ dedicated to keeping the nest clean. But ants achieve this familiar end result in a very different way to humans.
Human societies have centralized control. In other words, someone tells us what to do. Ants, on the other hand, have decentralized control and neither the queen nor any other ant directs work. Ant workers are the ultimate self-starters, following specific, but potentially flexible, rules in certain situations.
Chemical trails underpin much of this self-organisation. Foragers lay a mix of chemicals known as trail pheromone behind them as they walk. Other ants follow the trail and if they find food they reinforce it, laying more pheromone as they return to the nest. Stronger trails are more likely to be followed, so trails leading to food become progressively reinforced, while trails with no food at the end evaporate away.
This combination of positive feedback and evaporation produces an effective foraging system that is very good at finding the quickest routes to food. This simple guiding principle, and others like it, has provided some elegant solutions to the complex problems faced by engineers, computer scientists and businesses alike.
Getting your deliveries on time
WE’VE ALL WAITED for a package that didn’t turn up on time. And it seems that ants do a better job of delivering their parcels – or more specifically leafs – than your postman does. In finding the quickest way to food, ants are solving a routing problem. Businesses that need to deliver products while minimizing costs must also solve routing problems. Scientists have discovered that we can borrow the principles of ant pheromone trails to assist with our own ‘foraging problems’.
The best known routing conundrum is the ‘travelling salesman problem’ (TSP). The TSP seeks to find the shortest route between a number of different points and this becomes progressively harder as the number of points increases. However, simulations using the principles of pheromone foraging in ants, an approach dubbed Ant Colony Optimization, have been very successful in solving TSPs.
One example is Air Liquide, which supplies gas to a large number of customers across the USA, making this a particularly complex TSP. To solve it, the company uses a routing system based on Ant Colony Optimization, with trucks laying ‘virtual pheromone’ standing in for ants. Computers run through the night to calculate the most efficient routing solution for the next day’s deliveries, saving fuel and time.
Board your plane quicker
FINDING A SEAT on your flight seems to take forever when you’re stuck in a queue. So Southwest Airlines of the USA looked to ants for inspiration to speed things up. The airline allowed customers to choose seats as they boarded a plane, rather than assigning them a seat number in advance. But it wondered if the alternative was more efficient.
To test which was really the best boarding method, the company created a computer simulation that replaced humans trying to find a seat on a plane with ants. The virtual insects were given a simple set of rules, such as to find a seat when possible, to wait if the path was blocked, and to ask other ants to move if they were in the way. The ant-inspired simulation concluded that assigning seating beforehand was actually slower than giving passengers a choice on the aircraft.
Quicker computer networks
A FORAGING ANT colony has a network of trails along which it sends foraging ants to collect food and bring it back to the nest. By following simple rules the ants are able to use foraging trails to self-organize food collection.
But using this system is more than just following a trail of breadcrumbs. As well as finding food, the colony wants to ensure that it sends out the right number of foragers. If there are too many for the food available then the colony is wasting resources and risking lives. But too few means that the colony is not getting as much food as it could.
To solve this problem, researchers working on desert ants found that workers can use a very simple rule: the rate that workers leave the nest to find food depends on the rate that workers return to the nest with food. If there isn’t much food out there, the return rate of successful foragers is low and very few new foragers will leave, but a torrent of successful foragers signals a food-rich environment and the colony responds by sending out more workers. (The Behavior of Ants)
The ants’ foraging mechanism is almost identical to Transmission Control Protocol, or TCP – an algorithm used to avoid congestion on the internet. When a file is transferred it is broken into small packets and once each is received an acknowledgement (or ack) is sent back to the source. A high rate of acks shows that there is plenty of bandwidth and the transmission speed can be increased, much as a high rate of successful foragers returning to the nest means plentiful food.
Given that there are more than 11,000 species of ants that have evolved in many different ecological situations, researchers are keen to understand more about how they run their network operations. The hope is that the study of ant foraging networks will reveal other useful mechanisms that can inspire us in our own network management.
CUBESATS ARE MINIATURE satellites measuring just 10cm along each edge and are a relatively inexpensive way to do space research. Clyde Space, a Glasgow-based CubeSat manufacturer, has been investigating ant-inspired methods to build better satellites. With weight and space both at a premium, CubeSats need to be designed with the minimum amount of cabling. Just as ants use pheromones to find food, computer programs based on virtual ants laying virtual pheromones through a simulated CubeSat have created the most space-saving wiring solution.
Safer crowd control
VERY LARGE GROUPS of people in confined spaces can quickly turn into a life-threatening crush, especially if panic spreads. Constructing spaces with crowd safety built-in is an important part of modern architecture. Ant researchers studying Cuban leaf-cutting ants found that ‘crowds’ of ants contained in a space with two separate exits will tend to leave the space by both exits equally under normal conditions, but if a repellent chemical is added they will ‘panic’ and pile up around one exit, making it take longer to evacuate the area.
Similar panic-induced escape patterns have been found in theoretical simulations of human behaviour. Given this similarity, further investigation of the architecture of ant nests, and the ways in which different designs help ants move around within the nest could inspire human architects to come up with novel solutions for our own crowd control.
Forging fuel-saving routes through outer space
A SPACECRAFT CAN use the gravity of large bodies like planets to provide a ‘gravity assist’. By travelling on the right trajectory, the planet’s gravity increases the craft’s speed and changes its direction in such a way that, if you have done the mathematics correctly, it propels the craft towards its final destination. Gravity assists save on fuel and these celestial slingshots have been used to propel space probes like Voyager immense distances through space, Voyager 1 has travelled so far using this technique that it has now officially left the Solar System.
Although more complex than the traditional Earth-bound Travelling Salesman Problem, designing trajectories through space is still, at a fundamental level, a routing problem. However, to make use of gravity assists, the passage through space has to be combined with very accurate timing, which also makes this a scheduling issue. Researchers at the University of Strathclyde and the University of Glasgow have used the principles of pheromone foraging trails to construct a modified Ant Colony Optimization algorithm (set of instructions) that predicts routes through space. The algorithm removes the need to check all possible routes (a very time consuming process) and instead compiles the route incrementally, with each additional trajectory building on those the model has already ‘foraged’. Using this ant-inspired approach, they can predict the schedule and trajectories required to take advantage of multiple gravity assists far more rapidly than traditional methods.
Perhaps even the wise King Solomon couldn’t have predicted just how much we have come to learn from considering the ways of ants.