Debasis Basu

MODELING GENERALIZED COST OF TRAVEL FOR RURAL BUS USERS: A CASE STUDY

In order to formulate rational improvement proposals for rural bus services, it is necessary to understand how people value different attributes of travel. In this article, the disutilities of travel have been modeled based on stated choice data collected from trip makers traveling along a rural bus route in Midnapur District, West Bengal, India. Multinomal Logit Model is used to develop utility equations and the total disutility of travel is estimated in the form of generalized cost. The perceived values associated with in-vehicle travel time, service headway, and comfort level for the study route are estimated and found to be significant.

Stated Preference Approach for Valuation of Travel Time Displayed as Traffic Information on a VMS Board

A stated preference (SP) approach is explored for valuation of travel time displayed as traffic information on a variable message sign (VMS) board. Two types of traffic information (called as TI-I and TI-II) are investigated and the valuation is carried out separately for private car and taxi users. Instantaneous travel time is considered as traffic information in TI-I, while instantaneous travel time along with its variation level is considered as traffic information in TI-II. The SP (choice based) data are analyzed by multinomial logit and random parameter logit (RPL) models. In RPL models, the random parameters are assumed to follow constrained triangular distribution, where mean equals its spread. The decomposition of preference heterogeneity around the mean of random parameter is also investigated in RPL models. A variation is observed in the values of travel time under TI-I and TI-II, highlighting the role of the content of VMS based travel-time information. Both travel time as traffic information a...

Valuing attributes of enhanced traffic information: An experience in Kolkata

Abstract Most of the traffic information considers a single item like travel time or delay. In the present work, enhanced traffic information displaying instantaneous travel time and its variation from the previous interval to the present, is considered. An initial investigation is made on the effectiveness of such traffic information on route choice behavior of trip makers by valuation of attributes of the traffic information. Taking a case study of two urban corridors in the Kolkata metro city, India, the valuation is done separately for private car and taxi trip makers. The stated preference (choice based) data collected from trip makers are analyzed using both multinomial logit (MNL) and mixed logit (ML) modeling techniques. Assuming sparsely used constrained triangular distribution of random parameters, two different types of ML model are developed: one with independent choice sets and the other one by accounting heterogeneity around the mean of random parameter(s). Both family income and trip purpos...

Modeling stream speed in heterogeneous traffic environment using ANN‐lessons learnt

Abstract In order to model traffic stream speed resulting from complex interactions among different vehicle types in a heterogeneous/mixed traffic volume, an Artificial Neural Networks (ANN) approach is exploited. Two different categories of ANN model are attempted based on input vectors used. The performance of both categories of ANN model is evaluated using traditional evaluation framework. In addition, relevant logical test is carried out with both categories of ANN model. It is shown that selection of suitable input vectors and carrying out of relevant logical test are the two essential components for ANN model development process.

Value of travel time for home-based school tours in California

The behavioral value of travel time is an important factor for evaluating alternative transportation facility or service improvement projects based on road user-benefit approach. This article presents the behavioral value of travel time with reference to home-based school tour in California. Initially a single value of travel time is quantified from multinomial logit and nested logit (NL) model estimate. Later random parameter logit (RPL) model is employed by specifying a random component for the travel time attribute. The values of travel time emanating from RPL model estimate are quantified across student population by assuming different types of tent-shaped random distributions such as triangular and normal. In this study the value of travel time is investigated separately for two types of home-based school tour: grade school and college-university. Overall this study examines the effect of alternative logit model specifications on quantification of value of travel time. The study is demonstrated using household travel diary data of the state of California, which is revealed preference in nature.

Computational Analysis of a Single Server Queue with Batch Markovian Arrival and Exponential Single Working Vacation

In this paper, an infinite buffer queue with a single server and non-renewal batch arrival is studied. The service discipline is considered as exhaustive type under single exponential working vacation policy. Further, both the service times during the working vacation and normal busy period are assumed to be generally distributed random variables. It is also assumed that the service times and the arrival process are independent of each others. Moreover, it is accepted that at the end of an exponentially distributed working vacation, the first customer in the front of the queue is likely to receive service rate as per normal busy period service rate irrespective of received service in the working vacation period as the server shifts from working vacation mode to normal period mode. The system-length distributions at different epochs, such as post-departure and arbitrary epoch are obtained. The RG-factorization technique is applied to obtain the distribution of the system length at post-departure epoch. Henceforth, the system-length distribution at arbitrary epoch is determined by supplementary variable technique along with some simple algebraic manipulations. Some useful performance measures to be particular the mean system length of the model and the mean waiting time of an arbitrary customer in the system is discussed in the numerical section. Finally, some numerical results are presented for the model, in the form of the table and graphs. A possible application of the model in communication network is outlined in the paper.