Thesis etd-02182021-105105 |
Link copiato negli appunti
Thesis type
Tesi di laurea magistrale
Author
MASSIDDA, FABRIZIO
URN
etd-02182021-105105
Thesis title
A data-driven approach to individual carpooling benefit analysis
Department
INFORMATICA
Course of study
DATA SCIENCE AND BUSINESS INFORMATICS
Supervisors
relatore Prof. Nanni, Mirco
relatore Dott. Isfahanialamdari, Omid
relatore Dott. Isfahanialamdari, Omid
Keywords
- carpooling
- carpooling matching
- driver minimization
- individual mobility network
- mobilità
- mobility
- ride-sharing
- segmentazione di traiettorie
- traiettoria
- trajectory
- trajectory segmentation
- trip matching
Graduation session start date
05/03/2021
Availability
Withheld
Release date
05/03/2091
Summary
Un importante campo nel contesto dell'analisi dei dati di mobilità è quello del carpooling, che consiste nella condivisione di viaggi in automobile tra viaggiatori. Nonostante tale pratica abbia le potenzialità di migliorare le condizioni di vita, in particolare nelle aree urbane, arrecando inoltre benefici a livello ambientale ed economico, non si è ancora sviluppata in modo consistente. Ciò avviene per svariate ragioni, tra cui problemi organizzativi e di interazione. In questo lavoro viene presentato un approccio data-driven al carpooling, il quale sfrutta informazioni sulla mobilità personale, per esplorare le opportunità di carpooling tra individui, fornendo una sorta di analisi costi-benefici. Il metodo si concentra sulle necessità complessive degli individui in termini di mobilità, e affronta il problema di scegliere un insieme di conducenti ottimali per ogni utente passeggero osservato. L'approccio presentato viene testato su due dataset reali, contenenti informazioni circa veicoli dotati di GPS circolanti in Italia.
An important field in the context of mobility data analysis is that of carpooling, the sharing of car journeys among travelers. Despite the practice of carpooling may potentially enhance living conditions, in particular in urban areas, while also bringing environmental and economic benefits, it still struggles to develop in a solid way. This happens for multiple reasons, among which organization and interaction issues. In this work, we present a data-driven approach to carpooling that exploits information on personal mobility, to explore carpooling opportunities for individuals, providing a sort of cost-benefit analysis. The method focuses on the overall mobility needs of individuals and tackles the problem of choosing a set of optimal drivers for each passenger user observed.
The introduced approach is tested on two real datasets of GPS-equipped vehicles circulating in Italy.
An important field in the context of mobility data analysis is that of carpooling, the sharing of car journeys among travelers. Despite the practice of carpooling may potentially enhance living conditions, in particular in urban areas, while also bringing environmental and economic benefits, it still struggles to develop in a solid way. This happens for multiple reasons, among which organization and interaction issues. In this work, we present a data-driven approach to carpooling that exploits information on personal mobility, to explore carpooling opportunities for individuals, providing a sort of cost-benefit analysis. The method focuses on the overall mobility needs of individuals and tackles the problem of choosing a set of optimal drivers for each passenger user observed.
The introduced approach is tested on two real datasets of GPS-equipped vehicles circulating in Italy.
File
Nome file | Dimensione |
---|---|
Thesis not available for consultation. |