IT in the animal world: searching for food with ants and the TCP / IP protocol

A number of technologies that we use now were “invented” and implemented by natural selection and other evolutionary mechanisms millions of years ago. This, for example, ultrasound navigation (bats), sonars (whales), stun gun (electric eels), etc. As it turned out, nature developed millions of years ago the algorithms according to which the TCP / IP network protocol, already created by man, works. The methods of searching for food by ants coincide with these algorithms.

There are foragers in almost every ant colony of most species of ants, individuals that deliver food. At first glance, foragers randomly run around the nest, who are farther, who is closer, and there is no particular sense from their activities. But this is only at first glance. Despite the apparent randomness of their behavior, it is subject to certain laws.

The brightest case is the behavior of ant foragers of the species Pogonomyrmex barbatus. They almost never return to the colony without food - food. The frequency and number of "food raids" depends on the frequency and number of returning to the colony of foragers.
')
Many ant behavioral models have remained unchanged for many millions of years, so it can be said that these insects invented the basic principles of TCP / IP operation millions of years ago. According to the researcher Balaji Prabhakar , long time ago, ants learned about such functions as slow start of transmission (slow start) and interruption of transmission as a result of breaking the connection (time-out). As for the first case, when sending a large number of foragers (packets), the transmission continues, but with large intervals.

If there are no ants returning with food (there are no confirmation of the package delivery), the transfer is interrupted for a certain time. In the case of this type of ant, the delay is approximately 20 minutes. If during this period no ant has returned, the new foragers will not go anywhere, the nest is in standby mode. As soon as at least one forager with food appears , the delivery of food deliverymen resumes.

Route optimization: positive and negative feedback

Coming out of the nest, forager moves along an arbitrary trajectory. Behind him, he leaves an odorous trace (for other ants) using pheromones. The ant returns along its own path, and the trajectory of movement may be somewhat different (for example, going back with a load, he is forced to go around obstacles, which he just lightly conquered, like a mountain peak).

If the forager returns with a cargo, other ants "understand" that there, where the intelligence officer has just been, there is still food. Initially, other ants move along the original “paths”, but ultimately they lay a new, shortest route. The fact is that more ants have time to go through the short way per unit of time than the long way, which means that the pheromone trail of the short way will increase after some time, but the track of the long way, on the contrary, will weaken. In the end, the shortest route will remain, and the long paths will disappear due to evaporation of the feromon. Here there is one interesting nuance - if the pheromones did not evaporate, or evaporate more slowly than in fact, then it would not have been possible to pave the way for ants.


Reaper ant (Pogonomyrmex barbatus)

And so, since the system is based on positive feedback (when other individuals make the pheromone path more “noticeable”) and negative feedback (the pheromones evaporate, the path becomes less noticeable for other individuals), the ant movement trajectory self-organizes and eventually start moving along the shortest path.

Ants and robots

This year it became known about another study of ant colonies, the results of which were used in robotics. The fact is that scientists have long been surprised by the ability of insects to avoid traffic jams when digging moves. All of them somehow self-organize, the ants with a load pass the rest of the individuals.



As part of a new experiment, experts examined the behavior of 30 fire ants of the species Solenopsis invicta. In the process of creating a nest, they carried the load (soil particles) out. At first there were traffic jams in the tunnels, but over time, the ants learned to wait until the passage was cleared, and then carried their cargo without any problems.

Scientists studied the algorithm of ant behavior and used this algorithm to train small robots. Mechanisms were trained to transport cargo along an ant model, and robots began to work smoothly and without problems. The robots learned to let their fellows with the cargo pass and follow further along the planned route - without collisions and special delays. According to the authors of the study, the resulting algorithms can be useful for logistics, ro-mobiles, as well as building models for reducing traffic jams. Previously, only statistical models were used, but now it is quite possible to use the mechanisms that ants use.

IoT and ant colony separation

Another experiment, which was conducted several years ago, helped to understand why ants of one of the species decide to separate from their main colony and establish a new one (swarming of bees has approximately the same nature). So, as soon as the ants too often collide in the nest, the number of collisions reaches a critical value, then a number of workers "selects" the queen's heiress, picks it up and moves to a new place (not all species of these insects do this).

Scientists believe that the principle, which allows ants to determine the time when it is necessary to separate from the main colony, will help determine the reliability of the weather forecast based on data from sensors and sensors from a specific region. For example, sensors that exchange information with each other can transmit a certain token, and the central server, in addition to weather observation information, will receive information on the number of registrations of certain weather conditions in a certain place. The more such registrations, the more reliable the weather forecast, created on the basis of testimony.

The behavior model presented above for ants is also relevant for other social insects, including termites, bees, wasps and others. Further study of the behavior of social insects will make it possible to detect other "technologies" that they use. Some of them may well be useful to man. But in general, nature still has a huge number of secrets that are yet to be discovered.