Advanced Modeling for Transit Operations and Service Planning

342 Pages·2008·14.585 MB·342\342

by  Lam, William H.K., Bell, Michael G.H. (Eds.)

ADVANCED MODELING FOR
TRANSIT OPERATIONS AND
SERVICE PLANNING
edited by
Professor William H.K. Lam
Department of Civil and Structural Engineering
The Hong Kong Polytechnic University
Hong Kong, China
and
Professor Michael G.H. Bell
Centre for Transport Studies
Imperial College London
United Kingdom
United Kingdom North America Japan India Malaysia China Australasia
- - - - -
~
JAI Press is an imprint of Emerald Group Publishing Limited
Howard House, Wagon Lane, Bingley BD16 1WA, UK
First edition 2007
Copyright © 2008 Emerald Group Publishing Limited
Reprints and permission service
Contact: [email protected]
No part of this book may be reproduced, stored in a retrieval system, transmitted in any
form or by any means electronic, mechanical, photocopying, recording or otherwise
without either the prior written permission of the publisher or a licence permitting
restricted copying issued in the UK by The Copyright Licensing Agency and in the USA
by The Copyright Clearance Center. No responsibility is accepted for the accuracy of
information contained in the text, illustrations or advertisements. The opinions expressed
in these chapters are not necessarily those of the Editor or the publisher.
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library
ISBN: 978-0-08-044206-8
Awarded in recognition of
Emerald’s production
department’s adherence to
quality systems and processes
when preparing scholarly
journals for print
PREFACE
The idea for this book grew out of the organization of the Advanced Study Institute (ASI),
which was sponsored by the Croucher Foundation (http://www.croucher.org.hW) for the
dissemination of knowledge and the formation of international scientific contacts on advances
in modelling transit systems. While public transport (or transit) systems have arguably been in
existence much longer than road traffic systems, the mathematical analysis techniques so
necessary for the proper planning of transit operations have lagged far behind those for road
traffic systems. For example, the body of literature available on the design of schedules for
urban rail lines is miniscule in comparison to the literature on the coordination of traffic
signals along an urban road.
On the other hand, transit professionals appear to have disregarded most of the wealth of
insights that have been available in the literature for more than a decade. The literature on
transit assignment is a good example. However, public transport operators, particularly in
Hong Kong and Asia, are facing ever-greater pressure in competitive markets and transit
systems are congested. The need to estimate passenger demand, to monitor the performance
of individual services as well as the system as a whole, to support better planning and tighter
operations management, and for external reporting has increased. The optimization of transit
line frequencies and transit fares has become very important for operations and service
planning. Reliability and control issues are also critical in making transit systems more
efficient, supported by the introduction of Intelligent Transport Systems (ITS). As tightening
constraints raise serous questions about the cost-effectiveness of existing public transport
services, improvements which can be implemented in the short and long term are
continuously sought. Collectively, these pressures have focused attention on advanced
methods and new techniques for improving transit planning and operations.
In Hong Kong and other major cities in Asia, over 90% of people are using transit facilities
for their daily travel. The recent rapid development and deployment of ITS makes it possible
to improve the efficiency of transit operations. This book addresses the important and timely
problems of how to improve transit operations and service planning by making use of new
technologies and advanced modeling techniques. It will provide important references for
determining the outcomes of introducing these technologies and methods, and thus assist
transit professionals and scientists in resolving practical issues arising from the
implementation of ITS. This book appears to be the first devoted exclusively to the topic of
advanced modeling for transit operation and service planning.
This book consists of 12 chapters chosen to represent the broad base of contemporary themes
in modeling transit systems. Scholars from America, Europe and Asia have contributed their
knowledge to produce a unique compilation of recent developments in the field. Topics both
in theory and innovative applications to real world problems are included. The book covers
Transit Planning and Network Design, Transit Assignment Models and Solution Algorithms,
Simulation of Passenger Behaviors, Effects of ITS on Passenger Choices and Transit Service
Improvements, Modeling Multi-modal Transit and Urban Taxi Services.
Outline of the book contents:
Chapter 1 - Initial Planning for Urban Transit Systems
Chapter 2 - Public Transport Timetabling and Vehicle Scheduling
Chapter 3 - Designing Public Transport Network and Routes
Chapter 4 - Transit Path Choice and Assignment Model Approaches
Chapter 5 - Schedule-Based Transit Assignment Models
Chapter 6 - Frequency Based Transit Route Choice Models
Chapter 7 - Capacity Constrained Transit Assignment Models and Reliability Analysis
Chapter 8 - Dynasmart-IP: Dynamic Traffic Assignment Meso-Simulator for
Intermodal Networks
Chapter 9 - Modeling Competitive Multi-Modal Services
Chapter 10 - Modeling Urban Taxi Services: A Literature Survey and an Analytical Example
Chapter 11 - The Estimation of Origin-Destination Matrices in Transit Networks
Chapter 12 - Models for Optimizing Transit Fares
Special appreciation is extended to Elsevier Science Ltd. who made possible the publication
of all the contributions in the form of the present book in time to be available to participants
attending the AS1 workshop from gth to 13Ih December 2002 in Hong Kong. Professor Mike
Bell of Imperial College of Science, Technology & Medicine (U.K.) provided valuable
oversight and guidance in enhancing the quality of the book. His support during this effort has
been remarkable. Finally, I am thankful for the patience, availability, and dedication of the
editorial staff at Elsevier Science Ltd., particularly Julie Neden and Chris Pringle.
William H.K. Lam
Professor
Department of Civil and Structural Engineering
The Hong Kong Polytechnic University
Yuk Choi Road
Hung Hom, Kowloon
HONG KONG
Tel : [phone]; Fax : [phone]
E-mail: [email protected]
CONTENTS
Preface vii
Chapter 1 - Initial Planning for Urban Transit Systems 1
S. C. Wirasinghe
Chapter 2 - Public Transport Timetabling and Vehicle Scheduling 31
Avishai Ceder
Chapter 3 - Designing Public Transport Network and Routes 59
Avishai Ceder
Chapter 4 - Transit Path Choice and Assignment Model Approaches 93
Agostino NuzzoIo
Chapter 5 - Schedule-Based Transit Assignment Models 125
Agostino Nuzzolo
Chapter 6 - Frequency Based Transit Route Choice Models 165
Michael Florian
Chapter 7 - Capacity Constrained Transit Assignment Models and 181
Reliability Analysis
Michael G.H. Bell
Chapter 8 - Dynasmart-IP: Dynamic Traffic Assignment 201
Meso-Simulator for Intermodal Networks
Hani S. Mahmassani and Khaled F. Abdelghany
Chapter 9 - Modeling Competitive Multi-Modal Services 23 1
Hong K. Lo, C W Yip and K.H. Wan
Chapter 10 - Modeling Urban Taxi Services: A Literature Survey 257
and An Analytical Example
Hai Yang, Min Ye, Wilson H. Tang and S C. Wong
Chapter 11 - The Estimation of Origin-Destination Matrices 287
in Transit Networks
S. C. Wong and C. 0. Tong
Chapter 12 - Models for Optimizing Transit Fares 315
Jing Zhou and William H. K. Lam
CHAPTER 1
INITIALP LANNINFGO R URBANT RANSIT
SYSTEMS
S. C. Wirasinghe, Department of Civil Engineering, University of Calgary, Calgary,
Alberta T2N 1N 4 Canada
1. BACKGROUND
The current state of initial transit planning in many transit agencies could be described
at best as an art and at worst as a collection of ad-hoc rules. There are many reasons
for this situation. The complexity of the problems involved, the non-catastrophic
mode of functional failure associated with transit systems, the lack of trained
planners, political interference in detailed planning and the failure of people with
alternative planning tools to communicate their ideas to front-line planners have
contributed to the problem.
In the typical transit planning problem we are concerned with providing a good transit
service, which has a minimal environmental impact, at a reasonable cost to the transit
agency and to the users. A good level of service is provided by a transit service which
is reliable, easily accessible in time and space and provides a safe, fast and
comfortable ride at a reasonable price. The precise definition of the objectives to be
satisfied in providing the transit service, let alone their attainment, is a difficult task.
The problem is further complicated by the conflicting nature of the objectives.
2 Advanced Modeling for Transit Operations and Service Planning
A good estimate of the future demand for public transit is necessary to plan a transit
system. However, the demand, in addition to being a random quantity at any given
time, is also to some extent dependent on the type of transit system and its parameters.
This is another dilemma faced by planners.
A transit system is a failure if the objectives with which it was planned are not met to
a large degree. However, this type of functional failure, as opposed to engineering
failures, is not catastrophic since the system generally continues to function and to
satisfy the objectives to some degree. Further, errors in planning cannot be easily
pinpointed as the major cause of failure even when this is the case.
2. THEG ENERALP ROBLEM
Consider a city or a part thereof for which a public transit system is being planned.
The goal of such an exercise could be stated simply as the choice of the mix of transit
modes or technologies (e.g. Bus, Light Rail, etc.) and related optimal functional
designs, (routes, dispatching policies, etc.) for various areas of the city for different
time periods (peak, off-peak, etc.) that maximizes the expected utility to society.
However, the practical realization of the goal is not a simple matter.
In theory, the above problem could be converted into four related sub-problems:
(i) Determination of the set of relevant technologies or mixes of technologies.
(ii) Estimation of the present and future demand for transit given each of the
possible technologies (transit systems).
(iii) Optimal functional planning of each transit system for the related demand.
(iv) Choice of one of the transit systems as the ‘best’ one.
However, the sub-division does not provide us with four simple problems. Rather,
each one is in itself a complex problem.
2.1 Selection of Technologies
Various available technologies can be mixed in many ways for various areas of the
city, for different trip purposes and for different time periods. However, the number
of possible combinations is so large that it becomes prohibitive to carry out the rest of
the analysis (sub-problems ii to iv) for each technology-mix. So, a smaller perhaps
more relevant set has to be chosen based on speed and capacity considerations,
Initial Planning for Urban Transit Systems 3
compatibility with technology currently in use, environmental impact, geographical
constraints, etc. The interested reader is referred to the Canadian Transit Handbook
(1980), Gray and Hoe1 (1979), Vuchic (1981), and Parajuli & Wirasinghe (2001).
2.2 Demand Estimation
An extensive literature is available on the estimation of demand. A good introductory
work is that of Ortuzar and Willumsen (1994) . For further treatment of the modal-
split of demand see Domencich and McFadden (1975) and Daganzo (1979). In
general, one could proceed with the functional planning aspect under the assumption
that the demand for transit is given. However, in selected instances that assumption
can be released in favour of the one where the demand is a random quantity with
known mean and variance.
2.3 The Best System
The choice of one transit-system for implementation, out of several possible,
essentially boils down to a political decision. It is the planners’ duty, however, to
advise the decision-maker regarding the best choice. Bayesian Decision Theory
offers a rational approach by which the planner can take into account the several
options available, the uncertainty regarding demand, costs, etc. and a social utility
function. The reader is referred to Raiffa (1970) for an introduction to Decision
Theory and to Parajuli and Wirasinghe (2001) for an example of an application to
transit planning.
3. FUNCTIONAPLL ANNING
We are concerned with improving the present state of the functional planning of
transit and not with introducing a completely new methodology. An attempt can be
made to make the planning exercise more consistent by using analytical models based
on some of the more relevant and quantifiable factors that pertain to the problem at
hand. An analytical model can be optimised to obtain a theoretically sound ‘initial-
solution’ that can then be improved and ‘fine-tuned’ using all available hard and soft,
quantifiable and non-quantifiable constraints and other information. The ‘science’ of
transit planning can be considered to be the analytical modelling of a real system and
its optimisation, while the ‘art’ is the conversion of an optimal analytical solution into
a practical answer to a real, complex problem.
4 Advanced Modeling for Transit Operations and Service Planning
It should be emphasised from the beginning that several of the sociological, political
and geographic factors that affect a problem cannot be included easily in a model.
Thus the results obtained from optimising a model should not be taken too seriously
or as the “truth”. A model can, however, quantify some of the more relevant aspects,
serve as an initial solution and thus prevent one from proposing or implementing the
ridicule.
The following interrelated factors have to be considered when a transit system with a
specified mix of technologies is being planned for a given area.
3.1 Network
The network is the collection of routes for each technology for each time period. In
some cases, e.g. a variable route dial-a-ride service, it is sufficient to specify a zone-
of-operation.
3.2 Mode of Operation
The mode of operation is defined as the type of service offered on a particular route or
a network. For fixed routes, the service may be, for example, all-stop (local), few-
stops (express), non-stop, zone-stop (local or express in a zone and non-stop outside
the zone), or stop-on-call (e.g. dial-a-ride). For variable routes the service may be
non-stop, zone-stop, stop-on-call, etc.
3.3 Dispatching Policy and Fleet-Size
Given the mode(s) of operation, the rule according to which vehicles area dispatched
on the network routes is the dispatching policy. Essentially, the dispatch rate of
vehicles and the travel time on each route have to be specified. The network
dispatching policies, vehicle sizes and the fleet-size of each type of vehicle are
interrelated and should be considered more or less together. The detailed operational
aspects include the assignment of crews to vehicles and vehicles to routes.
3.4 Location of Transfer-Facilities
A transfer-facility is a location at which passengers can transfer from one mode to
another. Bus stops and rail stations are the most common examples. The location of
transfer-facilities is intimately related to the mode-of-operation. A terminal is a
transfer-facility at an end of a route.
Initial Planning for Urban Transit Systems 5
3.5 Location of Vehicles Garages
Normally, each mode will have one or more garages at which vehicles are parked and
maintenance is camed out. Occasionally, a terminal can be shared by modes with
some similarities and also serve as a terminal. The location of terminals is dependent
on fleet size, the network and to a large degree on the availability of land.
4. APPROACHTO AN ACCEPTABLAEP PROXIMATE
METHODOLOGY
Obviously, the five factors described above are interrelated and cannot be tackled in
complete isolation. However, each is a complex problem and their combination can
only be described as a ‘mess’.
The reason for the use of ad-hoc methods and rules-of-thumb in the functional
planning of transit is now fairly obvious. The fact of the matter is that the overall
problem is too complex to be formulated properly as a whole for solution even by
computer-based iterative methods. In any case, it would be nayve to write a complex
non-linear objective function with many non-linear constraints and optimise it with
respect to several variables using a massive computer algorithm, since even the output
from the most complex and largest possible computer program, would have to be
significantly altered by planners to allow for still unmodellable factors.
The basic premise of this chapter is that the ‘human’ planner reins supreme over
computing ‘machines’. The ‘block-box’ approach where a planner’s function is
simply to run local data with a ‘canned’ program and to accept the output as
inviolable is rejected.
The methodology proposed here is to break up the functional planning problem into
more or less independent sub-problems and to obtain approximately optimal
analytical solutions to the sub-problems. It is conceded that optimising parts of the
problem does not necessarily lead to the optimal solution to the combined problem.
The exercise of analytical modelling forces us to think formally about the problem at
hand and to isolate the critical factors. Thus it enhances our understanding of the
problem.