The connection between Internet algorithms and ant foraging

What is the difference between man and ant? The answer is that ants don't http://www.aliyun.com/zixun/aggregation/8.html "> Internet."

Compared with people who have creative thinking, ants seem to be just creatures that rely on instinct. But two scientists at Stanford University have discovered that ants are actually smarter than we think, though they don't invent the internet, but they know the algorithms on the Internet. On the Stanford University website, two scientists told the interesting findings.

When Deborah Gordon, a biology professor, observed Arizona, State's farm ants, he discovered their special behavior in foraging, so he found a computer scientist Balaji Prabhakar to explore the matter. Balaji Probhakar, though an expert on network file transfer, did not immediately realize how this was related to computers. But on the second day, suddenly came the inspiration, he suddenly thought that the behavior of ants and the Internet file transfer is actually very similar. "The algorithms that ants use to understand how much food is available are essentially the same algorithms used in TCP," says Prabhaker. ”

TCP, the Transmission Control Protocol, is one of the most important protocols on the Internet. It is used to manage data congestion on the Internet. It is because of the TCP protocol that our network can grow from the first dozens of nodes to today's billions of nodes. It works like this: A file is routed to endpoint B as a data source, and the file is decomposed into countless packets. When B receives a packet, sends a confirmation to a to indicate that the packet has been delivered.

The advantage of this is that data congestion can be avoided. If the B confirmation is slow, it proves that bandwidth is not enough, a can slow down the data transmission according to the situation, if B confirms fast, a will speed up the transmission. This process allows you to determine the bandwidth you can use and increase the transmission speed accordingly.

Gordon found that farmland ants behave very similar. They send foraging ants to feed. If there is more food, the ants will come back soon, and more ants will leave their nests to feed, and if the foraging ants return empty-handed, the ant colony's foraging behavior would be reduced or even halted.

In this case, Probkahar has compiled an algorithm to predict the foraging behavior of ants based on the amount of food, which Gordon has confirmed by experiments. They found that a TCP based algorithm almost accurately predicted the behavior of ants.

"The ants have discovered the algorithms we know well and they've been using it for millions of years." Prabhakar said.

They also found that the ant followed the two other rules of the TCP algorithm. One is called a slow boot, and at the beginning of the data transmission, a large number of packets are emitted to estimate the bandwidth;

Another is called timeout. When the link to the data transmission is interrupted or disturbed, the source stops the packet; Similarly, when the foraging ants do not return to their nests for more than 20 minutes, other foraging ants will not leave their nests.

Gordon argues that scientists have just come into contact with the surface parts of how ant colony behavior can help us design a networked system. There are now 11,000 kinds of ants, living in various environments, dealing with a variety of ecological problems, the way ants do things we do not expect, but may be used in computer systems. The capacity of each ant is limited, but the ant colony can accomplish complex tasks.

"So the ant algorithm must be simple, distributed and scalable, which is what we need to design a huge, artificial distributed system." "I think that when we learn more about how different types of ants regulate their behavior, we can find many useful examples of network algorithms," she said. ”

Ants and algorithms seem like incredible connections, but the wonders of nature often go beyond our imagination. The various algorithms and models discovered by man seem to be the creation of their own thinking, but they often find various examples in nature. This can not help but create a strange fantasy. Or, the world we live in is really an algorithm-driven machine, who knows?