DEPARTMENT OF MANAGEMENT STUDIES
INDIAN INSTITUTE OF SCIENCE
Ph.D. Thesis Colloquium of
Ms. Rupkatha Ghosh
[Research Supervisor: Dr. M. Mathirajan]
Date: 3rd December 2025 [Wednesday]
Time: 02:00 PM
Venue: Seminar Hall [Management Studies]
Title: “Development of Analytic Methods for a Class of Decision Problems in Urban Road Transport Organizations for Efficient Operation of City Buses”
Abstract:
Every Indian Urban Road Transport Organization (I-URTO) faces significant financial losses and increased CO2 emissions. Given these facts, this study considers the following 4 decision problems related to I-URTO operations to minimize the operational cost of city buses:
1. Allocation of City Buses to the Depots (ACBD) problem 2. Location of Depots (for opening new depots and closing some of the existing depots) and Allocation of Buses to Depots (LD-ACBD) problem 3. Problem on measuring the relative efficiencies of the I-URTO and developing/suggesting improvement areas for relatively inefficient I-URTO compared with efficient I-URTO 4. Problem on identifying and prioritizing motivating factors in the usage of city buses to increase revenue
The first research problem on ACBD is motivated by the significant non-revenue-generating distances, like Dead Kilometers (DK), which lead to higher operational costs and increased CO2 emissions. The analysis of the literature indicates that no existing study addresses the ACBD problem by simultaneously minimizing the associated DK Cost (DKC), Depot Operating Cost (DOC), and Environmental Cost (EC). For the ACBD problem considered in this study, an existing (0-1) ILP model is extended to minimize the following costs simultaneously: DKC, DOC, and EC. Due to the computational intractability of the (0-1) ILP model and to have very simple to understand and easy to implement efficient methodologies, in this study, heuristic algorithms are identified from the existing literature. Since the ACBD problem can be viewed as a special type of Transportation Problem (TP), 40 existing heuristic algorithms considering the criteria: (a) published in high-impact journals/conferences/ books, and (b) well-established computational experiments are identified to solve the ACBD problem. In addition, 4 new variants of existing heuristic algorithms are also proposed. Based on both empirical and statistical performance analyses w.r.t. the exact solution as the benchmark solution, considering randomly generated 360 large-scale ACBD problem instances, 3 heuristics: UPCM-TOCM, WUPCM1-TOCM, and VAM-TOCM, consistently outperform all others. These algorithms also result in 9-10% cost savings compared to the current practice method of I-URTO.
The second research problem addresses the integrated LD-ACBD problem, a strategic-level decision problem where expanding the city area, which increases the number of buses, requires new depots, while older depots in congested inner areas often need to be closed to avoid public inconvenience. Analysis of the existing literature indicates that very few studies considered the integrated LD-ACBD problem, particularly those that jointly consider the opening and closing of depots along with the ACBD problem. Further, existing studies on LD-ACBD problems have not considered either EC or DOC, or EC and DOC. To address this research gap, the (0-1) ILP model available in the literature is extended appropriately to include both costs. Due to computational intractability in obtaining an optimal solution, this study presents heuristic algorithms. The existing-Greedy Heuristic Algorithm (GHA) for the LD-ACBD problem selects the depot opening and closing combination with the minimum (FC – SC) and solves the ACBD problem considering the details on (a) increased number of buses and (b) decided depots for opening and closing along with remaining existing depots. However, GHA does not justify why it evaluates only one combination to address the LD-ACBD problem. This motivated us to consider each of the possible combinations of opening and closing depots based on the given data on opening and closing, along with each of the efficient heuristic alg orithms identified in the first research problem for solving the ACBD problem, which is part of the LD-ACBD problem. Based on empirical and statistical performance analyses, considering randomly generated 90 large-scale LD-ACBD problem instances, it is observed that (i) each of the 3 heuristic algorithms considered in this study and the existing-GHA for LD-ACBD problem yield the same efficient solution, and (ii) the efficient solution always corresponds to the lowest (FC – SC) combination. Thus, the LD-ACBD problem can be efficiently solved by simply selecting the combination with the minimum (FC – SC) and then applying any top-performing heuristic algorithm for the ACBD problem.
Though there are studies that have considered DEA and evaluated the relative efficiency of URTO, there are no studies that have stat istically justified and used input and output variables in DEA. To address this gap, in the third research problem, a nine-year dataset (2011-12 to 2019-20) for 10 I-URTO is compiled from government reports for identifying all possible sets of input and output variables. A structured statistical screening, including checks for normality, multicollinearity, stationarity, cointegration, and causality, is conducted, and finally, two Quantile Regression Models (QRM) are developed, considering input variables and output variables, respectively. The QRM identifies Labor Cost (LC) and Fuel Cost (FC) as statistically significant input variables, and Operating Revenue as the significant output variable. Using these input and output variables, both DEA models (CCR and BCC DEA Models) are developed and solved in R (Version 4.5.1). From the solution obtained from the DEA models, it appears that Pune Mahanagar Parivahan Mahamandal Ltd. (PMPML) consistently emerges as the benchmark I-URTO, while Delhi Transport Corporation (DTC) and West Bengal Transport Corporation (WBTC) show the lowest efficiencies. The analysis of the solution suggests that the reductions are required in LC and FC for the inefficient I-URTO to reach the efficiency frontier.
The objective of the fourth research problem is to identify and prioritize, from the user’s perspective, the factors that motivate the adoption of city buses in I-URTO. Though there are some existing studies in this context, there is no scientific justification for the number of factors considered in the existing studies. Furthermore, it appears that the prioritization of factors in existing studies is not considered! Based on these research gaps, a comprehensive and focused literature analysis on the identification of factors is carried out, and 30 factors are identified. Finally, these identified factors are reduced to 9 unique factors: Fare, Comfort, Reliability, Staff, Safety, Availability, Accessibility, Information, and Environment, as each of these 9 unique factors is used in different names by different authors. For prioritizing these 9 unique factors, the Best-Worst Method (BWM) is used by collecting the required data from 169 respondents (comprising 139 adopters of city buses and 30 non-adopters) through an exclusive questionnaire developed for it. Overall, the unique factors: Fare, Availability, and Accessibility are the top three motivating factors. Particularly, the analysis considering two perspectives: ‘adopters’ and ‘non-adopters’ of city buses’ indicates that fare is most critical for ‘adopters of city buses’, whereas accessibility and reliability are more influential factors for ‘non-adopters’.
Finally, the study presents the main contributions, including managerial implications for applying the proposed analytical solution methodologies to each of the four research problems, as well as the limitations of the study and future research directions.
ALL ARE WELCOME