Abstract :
[en] In the last two decades the development of mobile technology and the ease of access to an internet connection helped the consolidation of the sharing economy paradigm. This new way of purchasing goods and services differs from old traditional business models since it enables a shared use of resources in order to save money and generate profit. As an important player in the sharing economy, sharing mobility continues, nowadays, to shape urban mobility with the introduction of different modes of shared transport such as carsharing, bike sharing, ride sourcing, and other collective mobility services. Different stakeholders participate in the creation and exploitation of these new mobility services: governmental agencies, customers, and private companies. Each of them has a specific purpose that can affect and stir the benefit of sharing platforms. Focusing on the carsharing service, on the business side and on the user side, profit and customer satisfaction are usually the main goals even if, at times, both difficult to pursue together. Competition on today's landscape leaves little room for both established and less established businesses. Opportunities to increase corporate profit become scarcer and more refined systems to better manage carsharing operations are needed to guarantee commercial viability.
Evaluating business models for carsharing is no trivial task. Several methods are used for assessing the quality of changes in some operations or to evaluate possible approaches. Combinatorial and stochastic optimization are used to answer decision-making problems in the case of deterministic or uncertain problems. The shortcoming of these approaches is that they are limited at solving problems related to fleet management or service planning as it is more difficult to have an overview in which multiple properties (e.g., demography, territorial distribution, specificity of the fleet, ...) of both supply and demand are considered. This happens because car sharing is a highly complex service that has many interdependent factors. Given this complexity, a more favorable approach to estimate the demand for the service - together with all its peculiarities - and to help operators in the decision-making process, is the simulation one. This criterion allows the interaction of multiple factors which, through functional relationships between the decision-making parameters of the supplier, can introduce indicators to evaluate the quality of the solutions that cannot be easily derived analytically.
This dissertation focuses on a simulation-based approach that aims to create a decision support system for carsharing business. This decision support system aims to use demographic and land use data as input, once the provider's needs are known, and to return solutions regarding the optimization of the carsharing service. The development of this thesis is conceived from the point of view of the service provider, even if considerations regarding the equity of the various strategies proposed therein for the service customers constitute an integral and fundamental part of the construction of this system of support for decisions.
In this manuscript, we discuss the introduction of different dynamic pricing strategies that aim to increase the profit of the carsharing service, along with other indicators such as the number of bookings and utilization time of vehicles. By developing different price models, the introduction of dynamic prices based on the quantity of vehicles present in the station at the time of booking is evaluated and the output of the implementation of a dynamic price based on the time of the day is examined. In the first part of this thesis, we discuss how it is possible to evaluate the quality of a carsharing service from the point of view of its members, focusing on how different strategies generate or can reduce inequalities due to different wages or purchasing powers. Furthermore, using data collected by a car-sharing company operating in Germany and the United Kingdom, Oply, we implement these same strategies in a scenario calibrated with real data. Finally, we propose a methodology for calibrating carsharing scenarios in an agent-based environment.Moreover, we use these scenarios to demonstrate how it is possible, once there is complete knowledge of the demand and the status of the offer, to attribute a certain price to a single booking that maximizes the profit of the service.
The overall results show that the introduction of dynamic pricing strategies does not always benefit all segments of the population and that the goals of a carsharing company are not always compatible with those of its members. Furthermore, they show how it is possible to increase the profit of a carsharing company accordingly to its position on the market, whether it has a total knowledge of the territory or not, whether it is an established company or not yet fully established. As we will also see in the final chapter of this thesis, the product of this work does not consist only in a practical contribution aimed mostly at carsharing companies, but also in a scientific counterpart that outlines new research directions.
