Traffic Bottlenecks? Solved!

Modern as it sounds, traffic jams have been known to humanity for many centuries. It is a wonder, then, that the way traffic congestion is formed continues to pose a mathematical challenge to researchers to this day. Multiple models have been formulated to try to explain the origin of traffic jams, as well as predicting their appearance. In an article published in Nature Communications, Prof. Shlomo Havlin from Bar Ilan University, along with Prof. Hai-Jun Huang from Beihang University, present a novel analysis of traffic bottleneck dynamics in two Chinese metropolises. Using artificial intelligence tools, the researchers formulated a method for forecasting heavy traffic bottlenecks, based on their finding that the initial growth speed of a traffic jam is highly correlated with its maximal size. This enables an accurate, real-time prediction in the initial 15 minutes of congestion. Such updates on navigation apps save time, and may even prevent the traffic jam or at least moderate it.