An Analysis of the Impact of Self-Driving Cars on Traffic Conditions

The goal of this paper was to assess the effect of self driving cars on traffic conditions in the Greater Seattle Area using data from the Consortium for Mathematics and Its Applications in the 2017 Mathematical Contest in Modeling. We built a computational, agent based, mathematical model by which we could vary parameters such as number of lanes and capacity of cars. After polling a sufficient sample space, we ran our model and used curve fitting techniques to create functions to model the system. We used the model to calculate average speeds for various highways in Seattle. After creating our model, we adapted the computational model to allow for self driving cars to show increase in average car speed in various conditions. Our results show the increase in average speed with an increasing percentage of self driving cars in terms of increased average speed. The advantages of our model are the agent based aspects, which allow us to observe and model the system’s behavior. Using this data to interpolate surfaces allows for more analytic techniques as well. The computational model is also flexible enough to poll data for different traffic conditions in other cities. Authors’ Note: This paper was initially written for the Consortium for Mathematics and Its Applications’ (COMAP) 2017 Mathematical Contest in Modeling and received recognition as a Finalist paper top 10 out of 1527 teams working on the same problem. For reference, this problem was problem C and our team’s control number was 70320. Permission was obtained to publish this paper provided that proper credit is given to COMAP. As such, the entirety of the problem and all preliminary data is property of COMAP. E-mail contact: vinit.ranjan@duke.edu, junmo.ryang@duke.edu, k.zhang@duke.edu Copyright © SIAM Unauthorized reproduction of this article is prohibited