Computational aspects of carsharing planning without relocation operations

• Autor(a) principal: Cristiano Martins Monteiro
• Data de Publicação: 2022
• Tipo de documento: Tese
• Idioma: eng
• Título da fonte: Repositório Institucional da UFMG
• Texto Completo: http://hdl.handle.net/1843/52755; https://orcid.org/0000-0002-9655-3683
• topic: Low-cost mobility; Mixed-integer linear programming; Simulation; Computação – Teses; Mobilidade urbana – Teses; Programação linear – Teses; Simulação (Computadores) – Teses.

Commuting is a routine task of people living in urban areas. Shared mobility services aim to offer different options to this routine by providing better comfort and faster trips than conventional public transport means, along with avoiding the clients' costs of owning a private vehicle. A properly planned carsharing service can be attractive even for who owns and drives a private vehicle but would consider not owning it anymore if a cheaper and more sustainable transport mean is available. Low-cost carsharing rentals can be achieved by suitably positioning the fleet along the city and by making the most of the shared vehicles according with a previous selection of which subset of trip demands can be served. This previous selection would choose which demands have a combining origin and destination, allowing these clients to use a same vehicle but in different moments, not requiring the carsharing company to relocate the fleet among stations due to different demands along the day and week. This work contextualizes the operational and computational challenges in planning a carsharing service; proves the NP-Completeness of optimizing the locations for shared mobility stations; proposes a Mixed-Integer Linear Programming formulation for this original problem, and another Mixed-Integer Linear Programming formulation which yields good locations for stations; and applies a polynomial time linear formulation to simulate and compare the performance of three different carsharing business models according with historical mobility data from the São Paulo Metropolitan Area. Results show that it is possible to offer a profitable low-cost carsharing service without performing vehicle relocations. However, only a subset of trips are served and clients must be flexible enough to walk to get to an available vehicle nearby. Results also demonstrate that trips selected to be served are similar among the different business models; are concentrated on São Paulo's downtown region; are shorter than the average trip, but otherwise behave in a similar way as compared to the complete set of trips; and the lack of parking slots may be a risk to the carsharing company.