Too Much of a Good Thing? A Simulation Approach to Quantifying the Effects of Competition on Autonomous Taxi Networks

Abstract

As the United States’ population continues to grow, and high rents in cities force workers to live out of walking-distance from their workplace, more and more people choose to use the roads and highways that already seem packed to the limit. Vehicular accidents remain a leading cause of death, commute times keep rising, and the environment continues to suffer. At the same time, however, technology is improving in leaps and bounds, making a solution to the traffic problem not only viable but profitable, incentivizing many companies to begin research on autonomous vehicles. However, just as ride-hailing services did not meet lofty expectations of reducing traffic in cities like New York due to competition with other ride-hailing services and modes of transportation, competition may inhibit autonomous taxi fleets from resolving the problems plaguing the contemporary transit system. This thesis analyzes the effect of autonomous taxi providers’ market share in the state of New Jersey through simulation. Using a generated data set that mirrors the mobility demand of the population of New Jersey, and given level of service assumptions on such as departure delay and vehicle capacity, the taxi fleet required to serve a respective share of the market is generated. Analysis of generated fleets yields key metrics such as vehicle miles, average vehicle occupancy, and fleet size for each level of market penetration. These are compared against the current state of transportation in New Jersey, with the aim of providing evidence that, while competition does have a negative effect on the potential benefits of an autonomous taxi network, the overall result is positive.