Building last-mile connectivity to urban public transit systems: A decision sciences perspective

Guide(s)

Murthy, Rajluxmi V

Department

Decision Sciences

Area

Decision Sciences

University

Indian Institute of Management Bangalore

Place

Bangalore

Publication Date

3-31-2025

Year Awarded

March 2025

Year Completed

March 2025

Year Registered

June 2020

Abstract

Cities matter. They are the engine of the global economy and are already home to more than half the world’s population. Urban passenger mobility is an important area of research in sustainable transportation. For the urban commuter, connectivity at the last mile is a critical driver enabling regular usage of the dominant public transit system in a city. Last-mile connectivity refers to the challenge of reaching public transit stops, such as a metro station, from one’s source and then from the station to their final destination. This dissertation is a fresh attempt to investigate specific problems related to adoption and usage of modern public transit systems due to gaps in serving the passenger at the last mile. Methodologically, our approach applies a mix of quantitative models and tools from relevant fields of Decision Sciences such as computational geometry, graph theory, network optimization, and statistical modelling over robust secondary and primary data drawn from multiple publicly available sources. We revisit the last-mile problem as defined in extant transportation literature from the perspective of the current and potential passenger. Analysis of revealed preferences data from current passengers highlights foundational principles of quickness, affordability, and reliability of last-mile services as the elements which critically impact the usage of public transit. These principles form our basic premise for comprehensive fixed-route feeder subsystems providing last-mile access to broader public transit services. However, recent transportation literature has shown a tendency to focus on demand-responsive paradigms at the last mile, largely due to technological advances in the passenger transit ecosystem, providing us with a research gap. We address the opportunity by demonstrating the effective design of fixed-route feeder systems to solve the last-mile problem through a holistic approach motivated by passenger welfare, rooted in relevant research themes. We use spatial analytics and network optimization models to develop and implement a replicable framework to systematically quantify last-mile coverage, compute the components of the last-mile universe such as the last-mile stops and sub-zones, and design an optimal network structure for a fixed-route feeder subsystem that accommodates critical passenger and service provider requirements. We then conduct a conjoint analysis on stated preference data elicited from potential passengers through a specifically designed choice-based experiment. The analysis substantiates the nature of possible value propositions from such a fixed-route feeder service, which will induce a behavioral shift in passengers toward usage of public transport, and quantifies this impact. We consider the Metro Rail system in Bengaluru, India as a case study to illustrate empirical results within a geographical and administrative scope. Revealed preference data indicates that more than 70% of the current passengers walk the first/last-mile components of a metro journey due to the lack of effective feeder subsystems. However, only 47% of the current population is within walkable access of 1 km from the metro network, resulting in a loss of ridership which our framework addresses. The spatial tessellations-based approach computes the precise catchment areas of the stations, and identifies the last-mile universe comprising 245 sub-zones and stops on the city’s road network. We then use an Integer Linear Programming formulation to solve a network optimization problem which yields 63 independent optimally connected feeder networks, each having 4 or fewer routes that can be covered from the nearest metro station within 25 minutes, the maximum likely travel duration for a first or last mile leg. The combined multimodal system accommodates 93% of the population spanning 89% of area within the metropolitan region. Conjoint analysis of stated preferences data suggests that passengers are more sensitive to price changes at the last mile than for the metro ride. At least 15% of the decision of whether to use public transport for the journey depends on factors at the last-mile, affecting 77% of potentially competing journeys by hired personal transport. Our contribution to research and practice in the area of last-mile connectivity is multifold. We motivate a differentiated perspective on the last-mile problem in public transit and provide sustainable, passenger-centric solutions through a unique approach, making effective use of computational methodologies based on spatial tessellations, Voronoi diagrams, and network optimization formulations that have not been applied extensively in prior work to these problems. We also provide a quantitative assessment of the impact of last-mile value propositions on the primary travel mode choice, which has not been addressed in transportation research. Lastly, we have worked with data and proposed methodologies and frameworks which are replicable over geographies and administrative regions. The replicability is suitably demonstrated within the scope of the thesis with an implementation over a different city’s metro network (Delhi). Variations of the last-mile problem are ubiquitously encountered in passenger transport, and we envision that future research will leverage this work to address elements of last-mile connectivity that were beyond the scope discussed here.

Pagination

vii, 201p.

Copyright

Indian Institute of Management Bangalore

Document Type

Dissertation

DAC Chairperson

Murthy, Rajluxmi V

DAC Members

Chandrasekharan, Reshma; Prakhya, Srinivas

Type of Degree

Ph.D.

Relation

DIS-IIMB-FPM-P25-17

This document is currently not available here.

Share

COinS