Highway Research Record

Modeling Route Choice Behavior of Private Vehicle and Taxi Users: A case study

Duration: June 2004
Indian Institute of Technology, Kharagpur (S, I)

Scope and Objective
To develop utility model as well as generalized cost model for understanding the route choice behavior of private vehicle and taxi users. The case study includes travel from Howarh station to different parts of Kolkata via two routes: The route via Rabindra Setu (or Old Howarh Bridge) and the route via Vidyasagar Setu (or Second Hooghly Bridge). The scope of the present work includes:
(i) Design of Survey Experiment (Revealed Preference and Stated Preference)
(ii) Collection of Revealed and Stated Preference data
(iii) Development of utility equation for route choice behavior of taxi and private vehicle users
(iv) Development of Generalized Cost (GC) model for route choice behavior of taxi and private vehicle users

The proposed methodology includes design of Survey Experiment (Revealed Preference and Stated Preference). To understand the travel behaviors and route choice of the passengers, RP and SP Surveys are conducted. These surveys are carried out on-broad in trains in the stretch of South-Eastern Railway between Kharagpur and Howrah, and also at Howrah station parking area for car users. The attributes considered for the preparation of choice sets are Travel Time, Travel Cost, Transfer Time at Station, and Route Congestion Level. The methodology used for the work is summarized below:
Design of Experiments:
• Selection of survey method/Type of data: SP and RP data
• Preference elicitation method: Rating Based (Probability scale)
• Preparation of Choice sets: Fractional Factorial technique

Source of Data: Primary (Personal Interview) and Secondary information
Type of Model: Binomial Logit Model

Findings and Conclusions
Stated preference techniques are found to be instrumental for developing utility models and estimating perceived values associated with various attributes of travel for taxi and private vehicle users in Kolkata. The responses obtained from private vehicle and taxi users are also found to be consistent. The qualitative attribute ‘Road Congestion Level’ is found to be insignificant as the effect of road congestion is reflected in the travel time. Also, it is found that when transfer time and in-vehicle travel time are considered separately, the variable ‘transfer time’ does not become significant. Therefore, in the present work, the total travel time (i.e. transfer time and in-vehicle travel time together) is considered as one variable, and the perceived value associated with total travel time is estimated separately for private vehicle and taxi users with reference to the study route.