Behavior Modeling and Performance Assessment of E-Scooter Rental Mobility Systems in Discrete Event Simulation Framework

Abstract

Recently, E-Scooter (ES) rental systems have played a significant role in urban mobility and particularly by improving first and last mile connectivity. However, they are facing a critical challenge to ensure an adequate supply of E-scooters and in most areas to meet the growing demand. This paper presents a comprehensive study on the modelling of e-scooter rental systems using Stochastic Petri Nets (SPN) and discrete event simulation, highlighting their potential in capturing the dynamics of such systems. Furthermore, a repositioning strategy proposed for maintaining the balance and availability of supply across different locations based on SPN modelling. Attesting the reliability of a model requires incorporation of a performance evaluation algorithm, and analysis key performance metrics such as the average wait time and rate of rented scooters. Additionally, the model is used to assess the impact of several decision parameters, e.g., the battery range, and fleet size on the overall behavior and dynamic of an ES rental system. The results indicate that the repositioning strategy involved significantly reduces the number of unserved users and decreases the unbalanced situations in network. This work provides important insights into system implementations (either in strategic or operational level) while highlighting the effectiveness of SPN in assessing the performance of these dynamic and stochastic systems.