Title:Addressing the Free-Rider Problem in Public Transport Systems

Abstract:Public transport network constitutes for an indispensable part of a city by providing mobility services to the general masses. To improve ease of access and reduce infrastructural investments, public transport authorities often adopt proof of payment system. Such a system operates by eliminating ticket controls when boarding the vehicle and subjecting the travelers to random ticket checks by affiliated personnel (controllers). Although cost efficient, such a system promotes free-riders, who deliberately decide to evade fares for the transport service. A recent survey by the association of European transport, estimates hefty income losses due to fare evasion, highlighting that free-riding is a serious problem that needs immediate attention. To this end, we highlight the attack vectors which can be exploited by free-riders by analyzing the crowdsourced data about the control-locations. Next, we propose a framework to generate randomized control-location traces by using generative adversarial networks (GANs) in order to minimize the attack vectors. Finally, we propose metrics to evaluate such a system, quantified in terms of increased risk and higher probability of being subjected to control checks across the city.