Placement Optimization for Advertisement Dissemination in Smart City

This paper studies a promising application in Vehicular Cyber-Physical Systems (VCPS) called roadside advertisement dissemination. Its application involves three elements: the drivers in the vehicles, Roadside Access Points (RAPs), and shopkeepers. The shopkeeper wants to attract as many customers as possible by using RAPs to disseminate advertisements to the passing vehicles. Upon receiving an advertisement, the driver might detour towards the shop, depending on the detour distance. Given a fixed number of RAPs and the traffic distribution, our goal is to optimize the RAP placement for the shopkeeper to maximally attract potential customers. This application is a non-trivial extension of traditional coverage problems, the difference being that RAPs are used to cover the traffic flows. RAP placement algorithms pose complex trade-offs. If we place RAPs at locations that can provide small detour distances to attract more customers, these locations may not necessarily be located in heavy traffic regions. While heavy traffic regions cover more flows, they might cause large detour distances, making shopping less attractive to customers. To balance the above trade-off, bounded RAP placement algorithms are proposed with respect to submodular and non-submodular scenarios. Since real-world traffic distributions exhibit unique patterns, here we further consider the Manhattan grid scenario and propose improved solutions. Extensive real trace-driven experiments validate the competitive performances of the proposed algorithms.