| dc.description.abstract | Most of the large Indian cities are experiencing a high growth rate in vehicle population, leading to problems related to road congestion and air pollution. A strong public transport system can lead to better road utilization and thus can help contain these problems. Though bus transport is the principal mode of public transport in almost all the major Indian cities, its growth is less than other modes of transport. The city structure has been changing, while the bus route structure is being modified only in an ad-hoc fashion. Bus route network design is a key element in the operational planning process for urban bus services. It can be used to design a demand oriented service and to attract more passengers, apart from leading to efficiency.
Many of the suggested approaches for the design of a route network are based on unrealistic assumptions about the city structure. Some have taken the real city structure into consideration and have attempted to generate an improved solution. Since it is difficult to capture all the intricacies involved in the decision making process in a mathematical model, attempts to implement the solutions of these models rarely meet with success.
In this study, the approach adopted is that of a Decision Support System (DSS). This approach combines the analytical power of mathematical modeling with intuitive power of human decision makers, to result in an interactive decision making tool. A DSS software has been developed which includes a graphical user interface to support the human style of decision making. It also includes operations research models to help the decision maker in evaluating his actions, and in suggesting solutions.
Route evaluation model is the heart of the DSS. Routes can be evaluated if loads on all the route-links are known. Thus, the problem being addressed is to find the load on each route-link, arising as a result of the given demand and the route structure. This problem is also called public transport traffic assignment problem. An LP based evaluation model has been developed which minimizes the total users cost in a route system for the given fixed demand. User’s cost is defined as a weighted sum of traveling, waiting and transfer costs.
The model exploits a specific situation in Indian metropolitan cities, where buses often run over sitting capacity, to reduce the objective function from non-linear to a piece wise linear form. Further, to generate transfer between routes, the model uses the radial city pattern of Ahmedabad to its advantage, as most of the transfers take place in bus terminals located in central locations.
Since the LP model becomes very large for the real size problem, its solution takes a large time, apart from demanding a high level of computing resources. As a part of the DSS, the model is expected to be used repeatedly. Hence, the time taken by the model is critical. To address this concern, a Heuristic model has also been developed, which uses the same assumptions as in LP and approximates the LP solution iteratively. The Heuristic uses a greedy strategy for selection of O-D pairs. For each O-D pair, it uses the same cost function (as in LP model) for the allocation of demand to competing routes. The comparison of results suggests that while the Heuristic model gives a near optimal solution, it takes
Significantly lower time than LP. Further, the Heuristic model allows a greater maneuverability in generation of route transfers and thus improves upon the quality of the solution.
Apart from the evaluation model, tools have been developed to aid a user in designing new routes, or in improving the existing route structure. Some of these are:
i) Calculating Frequencies: This model recalculates frequencies of all the routes in order to achieve a target Peak Load Factor for all the routes. The algorithm resets frequencies of routes not meeting a pre-specified criteria and runs the evaluate model to get the loads again. This process is repeated till stability is achieved.
ii) Guided Route Design: This model provides active guidance in designing a new route or in improving an existing route. The model lets the user fix origin and destination points for a new route while giving information of demands remaining to be Satisfied for different nodes. Thereafter, it allows insertion of new nodes, while providing information on extra demand being catered by visiting a node and the deviation needed from the existing route.
The DSS has been customized for Ahmedabad Municipal Transport Service (AMTS) for the present study. An OD matrix, based on an earlier study was adopted which divides the city into 27 zones. Four different scenarios have been analyzed. These are; current (1997) and future (2001) demand for full day; and, current peak period and off-peak period demand. Separate route rationalizations have been carried out, to meet operator’s and passenger’s concerns.
The important findings are:
i) AMTS should decrease the number of routes being operated, while providing better frequencies on the reduced set of routes.
ii) The current services are capable of satisfying only 81% of the current demand.
iii) AMTS needs an addition of 225 vehicles to its current fleet in order to meet the demand of 2001.
iv) Only 63% of peak period demand is being Satisfied. It is suggested to take some targeted measures to improve the services during peak hours.
Methodologically, this study makes the following contributions:
i) Use of a DSS framework and heuristic approach, for design and evaluation of bus route networks, which are less explored areas.
ii) Modeling transfer passengers, specially suitable for cities like Ahmedabad, with radial pattern.
iii) Modeling standing passengers, a feature prevalent in Indian cities, by adopting a piece wise linear objective function.
iv) Providing insight into application of computer aided modeling techniques to Urban Bus Transportation in Indian Cities.
Such a DSS will be useful for urban bus operators in Indian cities, for carrying out minor route rationalization as well as major route design. This will contribute to decreasing their cost of operations, and will enable them to provide better service to passengers. Since the final choice is always with the human decision maker, the solution generated has a greater chance of acceptability. | en |
