Network Traffic
https://www.pcmag.com/reviews/waze
https://www.digitaltrends.com/mobile/waze-vs-google-maps/
https://support.google.com/waze/answer/6078702?hl=en
Recently in class, we’ve been discussing market traffic, modeling different trends and behaviors with cars, and the varying routes that are taken to get from point A to point B. Specifically, we’ve been using directed graphs with weighted nodes to optimize travel times.
For instance, in class we analyzed a graph with two nodes: A (the starting point) and B(the end point). Connecting these two nodes were two edges (thus meaning that there were 2 paths between). The top edge had a weight of (x/10)+10; where X represents the number of cars traveling along that route, whereas the bottom edge had a constant weight of 50. We were informed that there were 400 drivers, and tasked to derive the resulting nash equilibrium. In this case, the Nash Equilibrium was achieved when 200 drivers chose the top route, and 200 drivers chose the bottom route – resulting in a travel time of 50 for all drivers. Let’s add an additional driver to the mix; driver X. When X reaches node A, there are 200 drivers on both routes, meaning that if X takes the top route, it has a travel time of 50.1, and if X takes the bottom route, it has a travel time of 50. Even though these values are different; X has similar payoffs – and regardless of which route is ultimately taken, the travel time only differs slightly.
However, what if at a given moment in time, there were 300 drivers on the bottom path, and only 100 on the top? In this case, if X choses to take the bottom route, the travel time is 50, whereas the travel time for the top route is only 40.1. Other possible distributions yield even higher discrepancies. Consider this; when X reaches node A, there are 400 drivers on the bottom route, and 0 on the top; meaning that the bottom route has a travel time of 50, and the top route has a travel time of 30.1. Naturally, 30.1 < 50 so X would choose the top route. However, what if X did not know how many drivers were on each route? What if instead of traveling from one node to the next, X had to traverse an entire graph? It is ultimately these questions that enable us to bridge our theoretical lectures to the real world.
During the latter half of summer, I embarked on a series of weekly roadtrips where I ping-ponged between the Catskills (2 hours east of Ithaca) and Springs NY (Tip of Long Island). In total, I completed this drive 8 times total (4 times each way), and for each of these trips, I relied on Waze for navigation.
In the realm of navigational apps, Waze is regarded as being one of the elite. In 2021, PCMag called Waze “a drivers best friend” citing its stellar route optimization and real-time crowdsourced traffic and accident reports as “a cut above the competition’s”. Similarly, in 2021 digital trends reported that Waze was a “thorough option for drivers”, claiming that waze “excels at finding alternate routes around accidents and traffic jams and alerts drivers in advance of road-related incidents”. Since 2013, Waze has been a subsidiary of Google. On their specifications site, they explain how “the power of Waze is in your hands. By simply driving around with Waze open on your device, you share real-time information that translates into traffic conditions and roadstructure”. Using this data, Waze then computes relative amounts of Network Traffic on the roads, and finds the optimized path. Furthermore, to increase its accuracy, waze constantly compares its current data to past trends to best predict future traffic, in order to ensure that the path is always the most-optimized path. Waze’s intricate system of crowdsourcing and complex statistical algorithms are interesting applications of using Network Traffic to optimize tasks in our daily lives.