Ants Show Us How to Find Things Faster
Humans have made astounding progress as a species—essentially dominating the Earth—thanks to technological innovation. Interestingly enough, non-human organisms have their own form of technological innovation that they use to survive in the natural world. Their innovations involve physical or behavioral adaptations that have evolved over time to help them persist in their environment, like the sophisticated mouthparts of a mosquito and the climate-controlled mounds that termites build. Humans have taken inspiration from the most unique and practical adaptations observed in animals to engineer better, more sustainable solutions to our own problems—this is the scientific field of biomimicry. Ants have proven to be a master in the art of living efficiently, and so one of the most well-known examples of biomimicry belongs to them.
Finding food is something that ants have got down to a science. Consider the last time you laid a sandwich down next to you on a park bench, or your plate of watermelon on the grass at a picnic. Ants are guaranteed to show up within minutes. Computer scientists have attempted to mimic ants’ food-finding ability through a set of algorithms. The Ant Colony Optimization (ACO) algorithm is meant to explore solutions to problems visually by tracing the possible pathways between a starting point and a destination and finding the shortest pathway as quickly as possible. This algorithm is based on how ants have naturally evolved to track down and demolish any food source within sight.
The Nose Knows
Here’s how ant colonies find their meals in real life. Worker ants search for food sources by randomly exploring the area around their colony nest. Individual worker ants scout independently, but many ants are doing this at the same time. A colony can cover as much space as possible by avoiding other ants after one-time encounters. Ants can recognize other ants from their own colony by touching each other with their antennae, essentially picking up the other ant’s scent, which is produced by pheromones. Ants from the same colony have slight variations of a shared signature scent.
Ants also use their antennae to detect food sources. Most species of ants are blind so they depend on their antennae as their only sensory organs. Once a food source has been detected the ant removes whatever she can carry from the food source and finds her way back to the nest, all the while laying down a pheromone trail along her path. Hence, a treasure map is created. This pheromone trail can be detected by other worker ants searching for food, and followed to the source. These other ants will also lay down a pheromone trail from the food source back to the nest, reinforcing the previous ant’s trail.
A Self-Erasing Map
What makes this whole system genius is that the pheromone trails are time-sensitive—they dissipate within two minutes. The shortest trails are therefore the most likely to be maintained because they don’t start to dissipate as quickly and are easier to reinforce. As these shorter trails develop persistent and more intense pheromone scents, more worker ants are recruited to these trails. Once a food source has been completely removed, the pheromone trail dissipates. The treasure and the map to it disappear at the same time. Ants maximize their efficiency here because the shortest path is the one that is co-opted by the colony, which means only the closest food sources are sought out and brought back.
This cooperative behavior inspired the ACO algorithm, which takes the idea of pathway optimization even further. The ACO system consists of virtual trails and nodes, including starting and ending nodes. The probability that an ant will move from one node to another depends on the amount of pheromone on the trail, how desirable the trail is (which depends on the length of the trail) and parameters controlling how important each of these things is to the virtual ants. The amount of pheromone on a trail gets updated regularly according to the amount of pheromone being deposited by ants and how quickly it dissipates.
The original ACO algorithm was developed by Marco Dorigo for his PhD thesis in 1992. Many versions of the algorithm have since been created, but the most commonly used ones are the original Ant System created by Dorigo, Ant Colony System (ACS) developed by Gambardella and Dorigo and Max-Min Ant System (MMAS) developed by Hoos and Stützle.
The ACO algorithm was first put to use in solving the Travelling Salesman Problem, where a salesman needs to find the shortest route to a number of cities within a given area without passing through any city more than once, eventually returning to the city where he started. You can imagine how this algorithm could be useful for any number of problems, and so far we’ve successfully used it to improve vehicle routing, scheduling and data traffic control, among many other things. The potential for Ant Colony Optimization seems to be endless.
Ants outdo us when it comes to getting tasks done through their sheer abundance and efficient communication, but this may change in the future as the field of biomimicry advances. While the ant colony lifestyle may not suit us humans, we can definitely learn a thing or two from them. Imitating ants, and the rest of the natural world, may just lead us to our most advanced innovations.
Need a visual explanation for Ant Colony Optimization? Check out this infographic!