To solve a complex computational problem, Japanese scientists used a single-celled creature that, paradoxically, coped with it faster and easier than a powerful computer.
A group of researchers from the Keio University of Tokyo decided to use the amoeba to solve the so-called "traveling salesman problem", a problem known in computer science. Its essence is as follows: imagine that you are a seller traveling from city to city, offering your goods. You want to maximize your efficiency in order to make as much money as possible, and therefore you want to find the shortest route that would allow you to get from one point to another.
There is no universal formula for this. The only way to solve the problem is to calculate the length of each of the possible routes and choose the shortest one. In addition, as new cities are added to this route, the computational complexity increases exponentially. So, for 4 cities you need to build only 3 routes, and for 6 - already 360. And if your path involves visiting 10 or more cities, then the bill will go to millions.
Keio University's solution differs from all other algorithms developed by researchers. The reason for this is the single-celled creature Physarum polycephalum, namely its mucus. P. polycephalum itself is a very simple organism that can do two things: move towards food and hide from light. Millions of years of evolution have made it abnormally efficient in both of these processes.
For the experiment, the amoeba was placed in a chamber made of channels, with some food at the end of each. The unicellular organelle instinctively pulled the flagella in the direction of a specific channel - and thereby triggered an alarm that turned off the light. The scheme is simple: each channel is an analogue of the city from the problem, and the choice of one of them affects the probability that the light will go out in other channels. The further such a “city” is from the current point, the more often the light goes out in it.
From the outside, this may seem like a roundabout and not very clear way to solve the problem, but it has one important advantage: an amoeba, unlike a computer algorithm, does not need to calculate every single path. Instead, it reacts passively to current conditions, thereby being forced to choose the most profitable route. How many "cities" do not add on the way - the time to achieve the final result will not change from this.
Paradoxically, as a result, the simplest organism solves the algorithm faster than any modern computer. Scientists have not yet fully understood why this is happening. "The mechanism by which the amoeba manages to choose the shortest route every time remains a mystery," they admit. But if such single-celled creatures can be put into the service of science, this will help not only in solving hypothetical problems: perhaps with their help it will be possible to revise the approach not only to modern computing algorithms, but also to computer security systems.
Promotional video:
Vasily Makarov