Abstract Major transportation network companies (TNCs) have promised to shift to 100% electric vehicles (EVs) in the next two decades, which places an increasing need to investigate the issues of ride-hailing services provided by EVs. Existing studies that model the EV charging systems and the TNC service systems omit the influences of the charging costs (i.e., electricity rate, and value of waiting time) on driver supply and passenger demand, which results in inaccurate prediction of system dynamics. This study is the first attempt to understand the influence of the electricity rate on the demand/supply of ride-hailing services and its implications. We compute the charging cost as the sum of the electricity cost based on the charging volume and the values of the expected waiting time. Specifically, we construct a queueing model framework to calculate the expected waiting time with M/M/k/C and synchronized M/M/1 queues, which models the charging and ride-hailing service processes separately. The experiment results from a two-symmetric-unit network show that the system performance metrics, such as platform profit and ratios of passengers served, have decreasing trends with increasing electricity rates. These trends shift when electricity rates and wage/trip fare rates change simultaneously, indicating the TNC platforms are able to achieve high profits by adjusting wage/fare rates to handle changes in electricity rates. Similar performance trends are validated by increasing electricity rates on large-scale experiments based on real-world trip demand. We further undertake sensitivity analysis and conclude that as the passenger demand increases, the system’s performance metrics, such as platform profit and the percentage of served passengers, gradually converge, within the constraints of the number of EVs and their battery capacity; the usage frequencies of charging stations follow the Pareto Principle, where roughly 15% of stations could serve most of the charging demand.

Highlights • Design a queuing network to model the charging cost impacts on the EV taxi system. • Explore the trends of system performances on a small two-symmetric-unit network. • Test the results by simulating real-world trip demand in Manhattan, NYC. • ‘80/20 rule’ rule in charging station usages: 15% of stations serve most of the demand.