Ant colony behavior simplifies analysis of network communication
The beauty of information science is that the simplest of examples and experiments can be used to explain the intricacies of broad concepts. For one, graph theory can be explained by almost anything–in this case, an ant colony. In the article “Exploring Networks Efficiently” by Larry Hardesty, researchers at MIT conducted a study based on the theoretical prediction that “random walks” performed by ants can accurately provide a foundation for estimating population density. The researchers consider “random walks” as a superior research method to random sampling due to the experiment concluding that it derives an estimate faster than a random sampling would. Furthermore, the strategies employed in the study is applicable to any network or graph, especially to those that explain how members of a social network are connected.
In the experiment, the ant’s environment was simplified as a grid. Using graph theory, the researchers explained that each cell is a node and it shares edges solely with those cells immediately closest to it. The “random walk” is defined as the ant’s movement from one cell to another. Similar to the way a data scientist would need one individual to create a particular network of individuals, the researchers chose one ant to create a sampled data for population density. Their intuitive explanation was that population density could be explained by the amount of times the ants ran into each other in a given area. One of the researchers stated that, “what we’re doing is giving a rigorous analysis behind that intuition, and also saying that the estimate is a very good estimate, rather than some coarse estimate.” In the end, their estimates not only concluded that ants are very adept at inferring the concentration of other ants around them, but also how when one studies a network, they must start at one point in order to find other points to form a connection or linked network. This experiment has the capability to yield better procedures for studying network communication due to the fact that the method of measuring the ant’s movements correlates to how data scientists describe and measure how a social or communication network is connected. Just as the ants randomly traveled to adjacent cells, individuals are conditioned to branch out and form a network.http://news.mit.edu/2016/ant-colony-behavior-better-algorithms-network-communication-0713