A group of researchers from Keio University in Tokyo have shown that the amoeba can solve the famous "traveling salesman problem" with near-correct results. And the way she does it allows her, under certain conditions, to outrun the computer in computing speed. An extraordinarily intelligent creature called Physarum polycephalum.
The task of the traveling salesman is to calculate the shortest route between the base and several customers who need to go around and deliver the goods to them. It is distinguished by an exponential increase in complexity - with 4 buyers there are only 3 possible solutions, but with 6 buyers the number of options increases to 360. In general, the problem was solved in the 90s, and today you can get results even for millions of buyers, which is in demand in international logistics.
The amoeba Physarum polycephalum is always hungry and afraid of sunlight. Scientists have created a kind of labyrinth - there are 64 sectors in the container where food is located, and the amoeba itself is located in the center and must move to get to food. Above the labyrinth, a lighting system is mounted, which forms light "barriers" on the way to food. The neural network controls the lighting so as to leave a given number of targets in the shadow for the amoeba and highlight the paths between them the brighter the greater the distance. That is, the amoeba needs to find the shortest paths between objects in order to eat and avoid light.
It turned out that the microorganism almost always solves this problem perfectly accurately. And the speed of decision making increases linearly with the complexity of the problem, although in IT science the growth of computational operations in this case should grow exponentially. It is difficult to explain from the standpoint of science, so experiments continue. Now ordered test tubes with architecture to simulate tens of thousands of objects-feeders - it is interesting to observe the behavior of the amoeba with such an avalanche-like complication of the problem conditions.