Master Thesis: Lukas Mohs
Ride-sharing is a mode of transportation that increases the car and infrastructure utilization by grouping individuals of similar routes into joint trips. The approach serves as a pragmatic solution to the global overuse of resources, environmental threat, and congestion problems in metropolitan areas. Students specifically qualify as early adopters of ride-sharing systems due to their financial situation and participation in the sharing economy. We contribute to the research field of ride-sharing by developing a solution for this target group. Firstly, this thesis reviews the history as well as the state- of-the-art of ride-sharing. Secondly, it draws conclusions from empirical data retrieved from commuters to the Garching campus of Technical University of Munich. Thirdly, the matching of passengers with drivers is discussed and formalized as an integer linear program. The design and development of a Software as a Service ride-sharing platform is finally presented and evaluated based on the survey data. The thesis proves the applicability of ride-sharing to our research environment and provides a future strategy for the enhancement of the developed software. The developed solution is accessible at https://ridebee.de/.
The traffic situation in metropolitan areas around the world highlights the need for action. The economic damage of congestion in Germany is estimated to be e100 billion annually, which is approximately 3% of the gross domestic product (Köhler-Rama et al., 2018). Urbanization is a global mega trend and aggravates this problem (Burkert, 2015). Political measures such as the ban of diesel-powered vehicles in cities with a high fine dust concentration accentuate the need for a pragmatic solution (Morfeld et al., 2014).
The concept of ride-sharing already showed success in projects in the United States of America (USA). Hence, it qualifies as an appropriate solution to the given problem due to its easy realization and high potential elsewhere. Cars in Germany carry 1.4 passengers on an average trip and 1.2 on the commute to work (Umweltbundesamt, 2016). The increasing environmental awareness and rising sharing economy, increasing gas prices as well as the distribution of smartphones are main drivers that can leverage the concept of ride-sharing to address the acute mobility challenge. Doubling the occupancy of vehicles from one to two would already half the amount of vehicles on the road. Digitally matching individuals and supporting them in the process of traveling will most probably facilitate the share of vehicles on a daily basis.
1. What are the existing strategies and approaches to solve ride-sharing?
2. How to formalize the commuter matching problem for the TUM campus in Garching?
3. How to design and implement a platform for daily ride-sharing at the TUM campus in Garching?
Ride-sharing was the subject of many studies undertaken in the past. This field of research overlaps with the three scientific sectors: transportation systems, operations research, and Information Technology (IT). Transportation systems find the best use of the given infrastructure. Operations research can be defined as an approach of solving the optimum assignment of trips and identification of routes within a road network. IT is the technological facilitator that dynamically solves the assignment problem and integrates the solution into the real world.
Early ride-sharing schemes in the USA were described and analyzed in many scientific publications such as Burris and Stockton (2004), Burris and Winn (2006), Reno et al. (1989), Kelly (2007), or Haselkorn et al. (1995). Ferguson (1997), Chan and Shaheen (2012) provided a historical analysis and classification of different epochs and Furuhata et al. (2013) as well as Siddiqi and Buliung (2013) compared currently existing ride- sharing services.
Amey (2011); Amey et al. (2011) and Tucker (2016) analyzed the viability of ride-sharing for the Massachusetts Institute of Technology (MIT) campus and University of Milan also developed and studied a ride-sharing service for students (Bruglieri et al., 2011; Lue and Colorni, 2009). Correia and Viegas (2011) specifically studied social aspects by conducting a large survey in Lisbon. Early corporate ride-sharing solutions in Germany were identified and described by Handke and Jonuschat (2012) and Schäfer-Breede (1996). Arning et al. (2013) conducted a user study in Germany about ride-sharing system requirements.
General concepts regarding the emergence of the sharing economy, peer-to-peer mar- kets and online platforms as well as how they are designed and managed is described in the literature work of Sundararajan (2016), Choudary et al. (2017), and Shirky (2008). Dittrich and Hodel (2003) wrote their thesis about the technical development of a ride-sharing platform and Calvo et al. (2004) also dealt with the technical design of such a system.
The underlying problem of matching ride-sharers was addressed by Niels Agatz (Agatz et al., 2010, 2012, 2011), Calvo et al. (2004), Baldacci et al. (2004), and Stiglic et al. (2015).
However, the scientific investigation of ride-sharing opportunities and the appro- priate implementation is still demanded: Agatz et al. (2011, p.533) claimed that "the development of algorithms for optimally matching drivers and riders in real-time has not received attention from the transportation optimization community to date". Practically, ride-sharing systems in Germany also failed due to the lack of adminis- trative integration, technological robustness, and usability (Funke, 2006; Handke and Jonuschat, 2012). Graziotin (2013, p.5) analyzed existing ride-sharing solutions and concluded that an "open and extendable technological infrastructure" was missing in order to facilitate ride-sharing.
These findings emphasize the demand for a successful practical implementation of a high-technology solution that unifies the theory of optimal matching with user requirements gathered from previous studies.
By formalizing the commuter matching problem as a combinatorial optimization problem, the theoretical foundation for the algorithm can be laid that identifies the best matching solution out of the set of all feasible solutions. In addition to the best theoretical solution one has to adapt the matching procedure to two aspects: the dynamic nature of trip assignments as well as the selfish behavior of all agents.
To exemplary study the market segment of students and faculty commuters, empirical methods will be applied. A quantitative study will be based on survey data and analyzed to describe the behavior and attitude of this potential customer group. Furthermore, adjustments of the modelled problem could be undertaken based on the identified valuation of time and travel costs.
The descriptive research analyzes previous and the state-of-the art solutions to solve ride-sharing in other countries including the given circumstances. Further, market competitors in Germany are studied regarding their matching approach and technology. From this analysis, major requirements for platforms in that segment will be visible.
As part of the Software Engineering process, which will be used to create the commuter matching platform, subdisciplines like requirements engineering, software design and testing will be dealt with. In particular, personas and their specific use cases will be defined, the platform architecture will be modelled and afterwards implemented. Finally, the matching functionality will be tested by simulating several simultaneous platform users.
Name | Type | Size | Last Modification | Last Editor |
---|---|---|---|---|
Final-Presentation-Master-Thesis-Lukas-Mohs-final.pptx | 54,82 MB | 07.05.2018 | Manoj Mahabaleshwar (account disabled) | |
Lukas-Mohs-Kickoff-MA-04.12.17.pdf | 10,76 MB | 07.01.2018 | Lukas Mohs | |
Master-Thesis-Lukas-final.pdf | 42,00 MB | 07.05.2018 | Manoj Mahabaleshwar (account disabled) | |
Proposal-ver03.pdf | 72 KB | 04.11.2017 | Lukas Mohs |