People study ants to write algorithms

By viewing an ant colony as an intelligent system consisting of many individuals following simple rules, Belgian scientist Deneubourg was able to simulate them using algorithms.

The process of using robots to simulate the insect world continues to expand step by step. A few years ago, researcher Simon Garnier created robot ants that could follow electronic news scent trails . This path is created by a row of overhead projectors, which automatically track the movements of each robo-ant; Meanwhile, light sensors are installed in each robot's 'head', so it can follow the glowing trails of its fellow robots.

The ants in the movie Ant-man also move according to a news system. (Photo: Marvel).

Basically, just follow two simple rules - just wander around exploring until you find a 'trail' or 'food', and stick to the strongest trail it finds - eventually the The robot-ant has found the shortest route through a maze.

The move toward studying ants as mechanical rule-followers also went hand in hand with the growing realization that an ant colony could function as a single organizational system, just as a computer does. collection of individual circuits. This idea was famously demonstrated by Belgian researcher Jean-Louis Deneubourg in the 1970s.

One of his best known experiments required using two different bridges to connect an Argentine ant nest to a food source. The two bridges are identical in every way, the only difference is that one bridge is twice as long as the other. Initially, the ants chose between the two bridges at random, but over time, the vast majority of ants chose the shorter bridge, for the simple fact that their scent information accumulated there quickly. than.

The ants' system has the ability to self-regulate very cleverly - the shorter the path, the fresher the scent of news, attracting a larger number. The key point here: Individual ants may be stupid, but they have what Deneubourg calls a high level of 'collective intelligence' .

By viewing an ant colony as an intelligent system consisting of many individuals following simple rules, Deneubourg took another step forward: He discovered he could describe their movements using mathematical formulas. learning, which can then be used to create computer models.

The ant colony algorithm - in which an infinite number of routes are initially explored, then the best routes are enhanced while others are obscured - has since been used to improve the UK telecommunications network, designing design more efficient freight routes, to organize financial data, to better distribute supplies during disaster relief operations, and to plan factory jobs.

Scientists chose ant colonies as models for their algorithms because ants regularly adapt their designs and explore new methods; They not only find the most effective solution, but also have other backup plans.