PERFORMANCE ANALYSIS OF URBAN BUS TRANSIT SYSTEM

Abstract

Bus Transport is the principal mode of transport for most of the urban population in India due to its door-todoor accessibility, flexibility and cheapness in comparison to other modes of transport available for urban mobility. The rapid rise in ridership, growing cost of operation, clashing of interest of commuters and non-commuters and various other socio-economic and political reasons,cause an enormous problem for the transit management to operate the system efficiently and effectively. As most of the transit undertakings depend on government subsidy to cover up the deficit, it is imperative for a transit management to make ©ontinuous assessment and monitoring of system operation. This involves regular tabulation of selected performance measures with the primary purpose of keeping account of how well the bus system is operating. Such monitoring process can quickly detect problems and also help in identifying productive and unproductive routes. A wide spectrum of operating policy alternatives are available for implementation with a view to improving transit performance. A large number of researchers in the developed countries have formulated routing and scheduling models to solve their urban bus transit problems. But the problems faced by India and the developed countries are dissimilar in the context of increasing ridership with limited facili ties and resources in India vis-a-vis the effort of the ii operator in the developed countries to attract more ridership by providing better level of service. Hence a basic question arises whether the techniques adopted in those countries can be directly applied in Indian urban environment. Some researchers in India, however, have attempted to solve the routing and scheduling problems by developing computer simu lated models to replicate routing and operational character istics of bus transit in some Indian cities. But the simu lation models appear to be more complex for the operators and decision makers and the lack of computational facilities prevents them from implementing these models. It is increas ingly felt that there is a need to develop a performance evaluation model so that the likely impacts caused by alter native improvement measures on transit performance could be scientifically assessed before implementation of a particular option. In the absence of such an evaluation technique, the transit agencies, at present, implement the policy decisions on adhoc basie without studying the impacts on commuters and operators. In this study a methodology is developed to evaluate urban bus transit performance on route by route basis. Six operational policy variables are considered which encompass a wide spectrum of alternative measures available for transit improvement. These are • frequency, number of stops, fare, running speed, service population and route length. Analytical expressions are developed to relate these policy variables with service variables and output variables. The service ill variables considered are • average operating speed, average in-vehicle travel time per passenger, average walking time per passenger, average waiting time per passenger and the average trip time. The output variables are • ridership, revenue, vehicle-kilometres scheduled, vehicle-kilometres accomplished, vehicle-trips scheduled, vehicle-trips accom plished, passenger-kilometres, number of buses, operator costs, transit user time costs and total costs. Various transit performance measures are assessed from these output variables. The selected performance measures evaluated in this analysis are » operator cost per passenger, operator cost per passenger-kilometre, total cost per vehicle-kilometre, total cost per passenger, total cost per passenger-kilometre, revenue per vehicle-kilometre, user time cost per passenger, passengers per vehicle-kilometre, load factor, deficit per passenger, subsidy, economic efficiency, operating ratio and operational efficiency. The analytical expressions are programmed into a computer package written in FORTRAN IV and run on DEC 2050 at the Roorkee University Regional Computer Centre. The programme is flexible and broad-based and can analyse any number of routes and time periods of service with a wide range of policy alternatives considered for evaluation. The CPU time is not significant as a system of 5 routes can be evaluated in only 5 seconds. For calibration and validation of the model, an extensive amount of data has been collected by conducting IV on-board and off-board transit survey in three selected routes of Bangalore Transport Service, From the observed da.ta, a probability density function following a normal distribution has been obtained for passenger service time at stops and a negative exponential distribution for delay time at inter sections and missed vehicle-kilome tres . Since ridership predictions after a policy change are largely dependent upon the value of demand elasticities, the observed data has been further analyzed to evaluate the tran sit demand elasticities with respect to in-vehicle travel time, walking time, waiting time and fare. A demand function in the product form has been assumed and regression analysis has been performed to compute the demand elasticities. The impact of policy changes on output variables aofi performance measures is shown by drawing various trade-offs and summarizing in tabular forms. Calibration and validation of the model have been demonstrated for base case condition. A comparison of the model values of various output variables and performance measures with the actual recorded data from the transit office shows that the variations are within a limit of ten percent implying reliability of model prediction. The sensitive areas of the model are ahown so as to justify its reliability and to advise the model user about the input data which require extra care in obtaining high accuracy. The performance evaluation model developed in this study would be a very useful tool to transit managers and decision makers of small, medium end large sized cities to meke scientific assessment of various transit improvement measures before implementation.