scispace - formally typeset
Search or ask a question
Journal ArticleDOI

Passenger route guidance system for multi-modal transit networks

TL;DR: In this article, a multi-modal passenger route guidance system, called eFinder, is developed to aid travelers with their combined mode-route choices, which forms a platform for disseminating public transit information and should complement further development and use of public transport network by enabling travelers to make informed choices.
Abstract: In many public transport oriented cities in the world, especially Asian cities, the public transport system has been developed extensively, to the extent that it has become increasingly difficult to navigate. Although inter-modal transfers are common and often necessary, a complete set of the routes across transport modes is generally not presented in a form that is accessible to travelers, as each operator would only publish its own routes. Moreover, the common nonlinear fare tables together with inter-modal fare discounts pose challenges to travelers in deciding their best routes. This study develops a multi-modal passenger route guidance system, called eFinder, to aid travelers with their combined mode-route choices. We discuss the architecture and features of this system in this study. This system forms a platform for disseminating public transit information and should complement further development and use of the public transport network by enabling travelers to make informed choices.

Summary (3 min read)

1 INTRODUCTION

  • In many transit-oriented cites around the world, the majority of the population depends on the public transportation system (PTS) for their daily travels.
  • Even if one were equipped with perfect transit route and fare information, finding the optimal path would not be a simple task.
  • Recently, there begin to have a few traveler information systems developed for transit users.
  • With wireless communication technology (such as mobile phones), access is further extended to virtually any locations.
  • Section 2 introduces the system architecture of eFinder.

2 SYSTEM ARCHITECTURE

  • EFinder has a three-tier system architecture with trip planning and interactive GIS and ASR functions.
  • The middle tier connects and queries the route database according to the users’.
  • Based on origin and destination information collected through either the GIS or ASR interface, the application server determines the best transit stations or stops for boarding and alighting.
  • The final routing results are delivered to the users either via a GIS map or telephone messages.
  • The response time of producing the optimal paths per request, therefore, does not involve calculating the paths in an online manner, but merely needs to trace back the optimal paths that are already pre-calculated and stored in a tree structure in the route database.

3.1 Web-Based Geographical Information System (GIS) Interface

  • Passengers input their origins, destinations and preferences to the system through the Geographical Information System (GIS) map interface via the Internet or through the Automatic Speech Recognition (ASR) system via the telephone.
  • Moreover, the GIS facilitates the manipulation, analysis, and presentation of the spatial information.
  • The prototype system of eFinder employs Intergraph GeoMedia® Web Enterprise as the webbased GIS engine for the presentation of map and route results, and for the input of locations through the map interface.
  • Users navigate the map by clicking on it and using the directional arrows.
  • The optimal routes and corresponding stations are shown through the same map interface in the main frame.

3.2 Automatic Speech Recognition (ASR) System Interface

  • EFinder includes the server-based automatic speech recognition (ASR) technology to allow users query the system with their voice.
  • To enable voice queries, their prototype is equipped with the Intel's computer telephony (CT) card, Dialogic D/41 JCT-LS.
  • According to the Chinese phonology, each character consists of only one syllable, and each Chinese syllable is divided into two sound segments commonly known as the initial segment and the final segment.
  • Observations at each CDHMM state are modeled by a 20-mixture Gaussian distribution.
  • The CDHMMs are trained from a large telephone speech corpus collected by the Chinese University of Hong Kong (Ching, 1997).

4.1 SAM Network Structure

  • State augmented multimodal (SAM) network is a network model for passenger flows in a multi-modal network.
  • Its network structure combines the networks of different modes and intrinsically includes the multi-modal network characteristics such as sectional fares, interchange discounts and probable transfers.
  • EFinder employs the SAM network structure for the calculation of optimal paths.
  • Each node in the SAM network includes four variables to denote the location and transfer information in order to capture nonlinear cost functions and probable transfer rules.
  • Links in the SAM network are divided into two kinds, “directed in-vehicle links” and “transfer links”.

4.1.1 SAM Nodes

  • Contrasting from the conventional and simple way presenting nodes, four state variables are used to describe each SAM node, including: Location ( i ): this state variable corresponds to the identity (ID) of each bus stop or transit station.
  • This state variable is used for modeling modal transfers.
  • The notations of origin and destination nodes, where passengers enter and leave the network respectively, require special attention.
  • The middle two 0’s in these two cases do not carry any physical meaning other than specifying them as an origin or a destination.
  • The last state variable denotes the status of starting or ending a transfer link.

4.1.2 Transfer Rules

  • The set of links connecting the SAM nodes is divided into two subsets, “directed invehicle links” and “transfer links”.
  • These multi-modal paths consist of sequences of transfer links and directed in-vehicle links.
  • Undoubtedly, a traveler who chooses to take the subway on the first segment may arrive at his destination without any transfers, completing the transfer state sequence of 0-2-0.
  • Associated with each transfer state is a transport mode, for example, mode 1 (auto) is associated with transfer state 1 whereas mode 3 (bus) is associated with transfer state 4.
  • What kinds of transfers are probable obviously a function of the locality and its practices or habits.

4.2 Shortest Path Algorithm

  • To prevent excessive response time for calculating the optimal paths online, all-pairs shortest paths from stations to stations are pre-calculated offline and stored in the route database server.
  • The route database is updated periodically whenever the transit system is modified.
  • The shortest paths are stored with a tree structure through the notation of predecessors.
  • The process time is typically with 1-3 seconds.
  • Thus the SAM network structure prevents the occurrence of negative cost cycles, causing the problem of infinite path loops.

5 CONCLUDING REMARKS

  • The authors contend that parallel to the emergence of a sophisticated public transit system, developing a transit route guidance system is equally important so that the system can be fully utilized.
  • Linking up eFinder with these messaging methods is an important development direction.
  • Finally, if one considers the system not only as a way to offer transit route information to passengers, but a way to collect passenger preferences and origin-destination data, then it can be used to build up a passenger demand database to calibrate transit assignment models, such as the SAM model (Lo et al.
  • The system offers lots of room for further refinement and development.

Did you find this useful? Give us your feedback

Content maybe subject to copyright    Report

1
PASSENGER ROUTE GUIDANCE SYSTEM FOR
MULTI-MODAL TRANSIT NETWORKS
Hong K. LO, C. W. YIP
Civil Engineering Department, The Hong Kong University of Science and Technology,
Clear Water Bay, Hong Kong
cehklo@ust.hk
Brian MAK
Computer Science Department, The Hong Kong University of Science and Technology,
Clear Water Bay, Hong Kong
ABSTRACT
In many public transport oriented cities in the world, especially Asian cities, the public
transport system has been developed extensively, to the extent that it has become
increasingly difficult to navigate. Although inter-modal transfers are common and often
necessary, a complete set of the routes across transport modes is generally not presented
in a form that is accessible to travelers, as each operator would only publish its own
routes. Moreover, the common nonlinear fare tables together with inter-modal fare
discounts pose challenges to travelers in deciding their best routes. This study develops a
multi-modal passenger route guidance system, called eFinder, to aid travelers with their
combined mode-route choices. We discuss the architecture and features of this system in
this study. This system forms a platform for disseminating public transit information and
should complement further development and use of the public transport network by
enabling travelers to make informed choices.

2
1 INTRODUCTION
In many transit-oriented cites around the world, the majority of the population depends
on the public transportation system (PTS) for their daily travels. For example, the PTS in
Hong Kong carries over 90% of the 11 million daily trips. As the PTS is being expanded
and further developed, as is typically the case, the system is becoming increasingly
difficult to navigate. To assist passengers in locating and using their services, transit
operators often set up web sites to publish their routes. However, due the competitive
nature of privately operated services, none of the operators would publicize the routes of
other operators, rendering such information to be piecemeal in nature, short of providing
useful route guidance information across modes or across operators. The need of a
common platform to provide routing information across modes and operators is apparent.
Even if one were equipped with perfect transit route and fare information, finding the
optimal path would not be a simple task. Most transit fare structures in practice are zonal
or origin-destination (OD) based, not directly proportional to the travel distance or travel
time, or non-additive in nature. Moreover, the increasingly common usage of electronic
smart cards for transit fare payment opens up innovative fare structures that are not
possible or otherwise difficult to implement with the traditional coin-based system. Some
examples include collaborative fare discount schemes for certain types of transfers
between routes and modes, as are widely practiced in Hong Kong. All these add to the
complexity of the fare structure in the multi-operator, multi-modal PTS.
Parallel to the emergence of a sophisticated PTS (for example, Chien and Spasovic,
2002), developing a transit route guidance system is equally important so that the PTS
can be utilized to its full extent. Such a multimodal route guidance system not only
allows travelers to find their ways to destinations, but more importantly, do so while well
aware of the alternatives and choose accordingly, such as minimum cost, time, or number
of transfers. Responding to this need, this study develops a transit route guidance system,
called eFinder, to facilitate passengers’ navigating the sophisticated multi-modal PTS.
Most route guidance systems or advanced traveler information systems (ATIS) are
developed for private car driving. Recently, there begin to have a few traveler
information systems developed for transit users. For example, Cashin et al (2002)
discussed the “Los Angeles Smart Traveler (SMART TRAVELER) Project” deployed by
the California Department of Transportation (Caltrans). The system provides telephone
and Internet information of traffic and transit services. There are also transit itinerary
planners developed for California, as depicted in the SMART TRAVELER
(http://www.dot.ca.gov/caltrans511/
), such as “OTIS – trip itinerary” of San Diego (San
Diego Metropolitan Transit System, 2004) and “511 Take Transit Trip Planner” of San
Francisco (Metropolitan Transportation Commission and San Francisco Bay Area
Transportation Partners, 2004). Travelers obtain transit services information for their trips
by entering their origins, destinations, and times of departure or arrival. Other examples

3
of itinerary planners available on internet include TransiTrips (TransiTrips, 2004),
TriMet (TriMet, 2004), and VIS’s Personnal Trip Planner (Metropolitan Transit, 2003).
The emergence of these transit information services shows the increasing demand for
such information. In Hong Kong, some transit operators provide route search services in
their own web pages, such as “Point to Point Route Search” of Kowloon Motor Bus
(http://www.kmb.hk
) and “Bus Route Search” of Citybus (http://www.citybus.com.hk).
However, they cover only their own routes but not the whole Hong Kong transit network.
Transit information services typically involve a route search algorithm that works with
multimodal networks. This route search algorithm to a great extent defines the
performance and quality of the routing information provided. Route searching in
multimodal transit networks involves procedures more complicated than simply finding
the shortest distance or time paths, which may involve the use of hyperpaths to handle the
common line problem and waiting time of transit routes (Nguyen and Pallottino, 1988;
Spiess and Florian, 1989; De Cea and Fernandez, 1993; Wu et al., 1994). In other
situations, with known schedule information, the search procedure may involve schedule-
based networks (Nguyen et al., 2001; Nuzzolo et al., 2001; Tong et al., 2001; Poon et al.,
2004).
As for the system developed in this study, eFinder takes advantage of the State
Augmented Multimodal (SAM) network structure technique (Lo et al. 2001, 2002, 2003;
Lo and Yip, 2001; Yip 2001) to handle complex fare and modal transfer configurations
by working with frequency-based transit services. Moreover, most existing such systems
only work with routes that include only one transfer; eFinder can handle trips with
unlimited number of transfers. On the other hand, it can also be set to find routes that do
not exceed the number of transfers as specified by the users. Subject to the travelers’
preferences: minimum fare, minimum time, or minimum generalized cost, eFinder
determines and advises the travelers with the most appropriate routes across modes and
across operators. Users of eFinder input their origins, destinations, and preferences
through either the Geographical Information System (GIS) interface via the Internet or
through the Automatic Speech Recognition (ASR) system via both fixed or mobile
telephones. The routing results are fed back to the users via these two interfaces.
When combined with the Internet, GIS multiplies its capability tremendously by allowing
access to the system from any connected terminals. With wireless communication
technology (such as mobile phones), access is further extended to virtually any locations.
Peng and Huang (2000) categorized online public transport information system according
to their content levels and functionalities, ranging from static to real time information,
and from text browsing to interactive map-based and online transactions. On the other
hand, speech is one of the most natural and efficient communication means. While the
mobile phone is a common communication device, using mobile phones for information
retrieval becomes a very attractive proposition. The system architecture of eFinder
includes both GIS via internet and ASR via mobile phone as input and output interfaces.

4
The outline of this paper is as follows. Section 2 introduces the system architecture of
eFinder. Section 3 presents the interface of the system. Section 4 describes the analytical
component of eFinder system and the SAM network technique. Section 5 contains some
concluding remarks and further system developments.
2 SYSTEM ARCHITECTURE
eFinder has a three-tier system architecture with trip planning and interactive GIS and
ASR functions. The system architecture is summarized in
Figure 1. The client tier provides interfaces for users to input their origins and
destinations and for the routing results to be presented. eFinder provides the web-based
GIS interface and phone-based ASR interface. For the web-based GIS interface, clients
navigate the map and input their origins and destinations by entering the addresses or
clicking on the map interactively. On the other hand, the ASR system allows the system
to identify the clients’ locations in a few dialogues. The results of route search are also
presented through the interfaces.
The middle tier connects and queries the route database according to the users’ OD inputs
and composes the routing results. The middle tier is the kernel of the system. Based on
origin and destination information collected through either the GIS or ASR interface, the
application server determines the best transit stations or stops for boarding and alighting.
It then queries the route database for the optimal routes and converts the results to a
format understandable to the users. The final routing results are delivered to the users
either via a GIS map or telephone messages.
The database tier mainly consists of the route database that stores the all-pairs optimal
paths according to different criteria including the shortest time, cheapest, or the balanced
choice. The response time of producing the optimal paths per request, therefore, does not
involve calculating the paths in an online manner, but merely needs to trace back the
optimal paths that are already pre-calculated and stored in a tree structure in the route
database. This setup substantially reduces the response time to produce the optimal paths.
The all-pairs optimal paths are pre-calculated offline by the Floyd-Warshall algorithm
(Cormen et al., 1990) based on the SAM network structure, as explained in Section 4. As
the transit route and fare information do not change from minute to minute, eFinder only
updates the route database periodically, say on the order of days or weeks or whenever
the transit system is modified.
The transportation information system (TIS) contains the up-to-date transit system
information including transit schedule, station location, transit fare, travel time, etc. The
TIS is typically maintained by the transport department or to be obtained individually
from each of the transit operators and composed together. Other than calculating the
optimal paths, the analytical model transforms the transit route, fare, and transfer

5
information from the TIS to the SAM network structure for the optimal path calculation.
3 USER INTERFACES
3.1 Web-Based Geographical Information System (GIS) Interface
Passengers input their origins, destinations and preferences to the system through the
Geographical Information System (GIS) map interface via the Internet or through the
Automatic Speech Recognition (ASR) system via the telephone. Moreover, the GIS
facilitates the manipulation, analysis, and presentation of the spatial information. The
prototype system of eFinder employs Intergraph GeoMedia® Web Enterprise as the web-
based GIS engine for the presentation of map and route results, and for the input of
locations through the map interface.
The Graphical User Interface (GUI) of eFinder is shown in
Figure 2. The main frame shows an interactive map. Users navigate the map by clicking
on it and using the directional arrows. They can also zoom in the map at various levels of
details and select their locations of interest by clicking on the map. The frame on the left
hand side provides users another interface to input their origins and destinations
hierarchically, by area, district, and finally address or landmark. Users then select their
preferences: fastest, cheapest, or eFinder (i.e., minimum generalized cost considering the
combination of fare, travel time, transfer penalties, etc.) routes.
The optimal routes and corresponding stations are shown through the same map interface
in the main frame. By gradually zooming in, one gets to the details of showing the exact
stations for boarding, alighting, or transfer (see Figure 3). The bottom frame shows the
detailed information of the resulting routes including the suggested modes, routes, travel
time and fare of each segment, and the transfer details such as modal change locations,
frequency of mode being transferred to, and any applicable interchange fare discounts,
etc.
3.2 Automatic Speech Recognition (ASR) System Interface
eFinder includes the server-based automatic speech recognition (ASR) technology to
allow users query the system with their voice. Users simply tell the system any landmarks
that are closest to their locations and destinations, then the system will process the query
and answer with the routing results including the modes, routes, fare, the station locations,
transfers if any, etc.
Voice query is particularly useful when a computer is not readily accessible especially
while en-route. On the other hand, telephones or mobile phones are ubiquitous. Moreover,
it is more natural to interact with eFinder by voice. Speech recognition will be
accomplished on the server side, thus relieving a user from expensive computers, smart

Citations
More filters
Journal ArticleDOI
TL;DR: The objective of this paper is to present a new formulation and an algorithm for solving the itinerary planning problem, i.e., determination of the itineraries that lexicographically optimizes a set of criteria while departing from the origin and arriving at the destination within specified time windows.
Abstract: The itinerary planning problem in an urban public transport system constitutes a common routing and scheduling decision faced by travelers. The objective of this paper is to present a new formulation and an algorithm for solving the itinerary planning problem, i.e., determination of the itinerary that lexicographically optimizes a set of criteria (i.e., total travel time, number of transfers, and total walking and waiting time) while departing from the origin and arriving at the destination within specified time windows. Based on the proposed formulation, the itinerary planning problem is expressed as a shortest path problem in a multimodal time-schedule network with time windows and time-dependent travel times. A dynamic programming-based algorithm has been developed for the solution of the emerging problem. The special case of the problem involving a mandatory visit at an intermediate stop within a given time window is formulated as two nested itinerary planning problems which are solved by the aforementioned algorithm. The proposed algorithm has been integrated in a Web-based journey planning system, whereas its performance has been assessed by solving real-life itinerary planning problems defined on the Athens urban public transport network, providing fast and accurate solutions.

153 citations

Journal ArticleDOI
TL;DR: A multimodal trip planning system for multiple modes: driving, transit, and driving-parking-then-transit, which considers the real-time transit arrival time, which is estimated by a prediction model.

30 citations

Journal ArticleDOI
TL;DR: This paper treats with K -shortest viable path problem in a transportation network including multiple modes, and finds that GASA is appropriate for small networks and BACS has better performance in median and large-scale networks.

26 citations

Journal ArticleDOI
TL;DR: A bus line optimization model based on bus and metro integration is constructed based on the co-opetition coefficient and Changsha Metro Line Number 2 is chosen as a case study to verify the optimization model.
Abstract: Metros are usually built and added on the basis of a completed bus network in Chinese cities. After the metro construction, it is faced with the problem of how to adjust and optimize the original bus lines based on the new metro system. This research mainly proposes a bus line optimization method based on bus and metro integration. In the consideration of the geographical space, the cooperation and competition relationship between bus and metro lines is qualitatively introduced according to the geographical location and service range of metro (800 m radius) and bus (500 m radius) stations. The competition and cooperation indexes are applied to define the co-opetition relationship between bus and metro lines. The bus line optimization model is constructed based on the co-opetition coefficient and Changsha Metro Line Number 2 is chosen as a case study to verify the optimization model. The results show that the positive competition, efficient cooperation, and travel efficiency between metro and bus has been significantly enhanced after optimization. Moreover, this paper provides a reasonable reference for public transport network planning and resource allocation.

12 citations

Journal ArticleDOI
TL;DR: In this article, the authors investigated the effect of indifference thresholds on transit fare and frequency schemes in a bi-modal network with the assumption of indifference threshold-based travelers' mode choice behavior.
Abstract: Summary This paper investigates public transit service (fare and frequency) operation strategies in a bi-modal network with assumption of indifference thresholds-based travelers' mode choice behavior. Under such behavior, users would switch to a new mode only if its utility is larger than the utility of current mode plus a threshold. The concept of indifference thresholds-based bi-modal equilibrium (ITBE) and the properties of the ITBE solution are explicitly proposed. Considering transit operator's different economic objectives (profit-maximizing, no-deficit and total system cost-minimizing), the effect of indifference thresholds on transit fare and frequency schemes is studied. Some numerical experiments are accompanied to verify the theoretical results. Copyright © 2016 John Wiley & Sons, Ltd.

11 citations

References
More filters
Journal ArticleDOI
Lawrence R. Rabiner1
01 Feb 1989
TL;DR: In this paper, the authors provide an overview of the basic theory of hidden Markov models (HMMs) as originated by L.E. Baum and T. Petrie (1966) and give practical details on methods of implementation of the theory along with a description of selected applications of HMMs to distinct problems in speech recognition.
Abstract: This tutorial provides an overview of the basic theory of hidden Markov models (HMMs) as originated by L.E. Baum and T. Petrie (1966) and gives practical details on methods of implementation of the theory along with a description of selected applications of the theory to distinct problems in speech recognition. Results from a number of original sources are combined to provide a single source of acquiring the background required to pursue further this area of research. The author first reviews the theory of discrete Markov chains and shows how the concept of hidden states, where the observation is a probabilistic function of the state, can be used effectively. The theory is illustrated with two simple examples, namely coin-tossing, and the classic balls-in-urns system. Three fundamental problems of HMMs are noted and several practical techniques for solving these problems are given. The various types of HMMs that have been studied, including ergodic as well as left-right models, are described. >

21,819 citations

Book
01 Jan 1990
TL;DR: The updated new edition of the classic Introduction to Algorithms is intended primarily for use in undergraduate or graduate courses in algorithms or data structures and presents a rich variety of algorithms and covers them in considerable depth while making their design and analysis accessible to all levels of readers.
Abstract: From the Publisher: The updated new edition of the classic Introduction to Algorithms is intended primarily for use in undergraduate or graduate courses in algorithms or data structures. Like the first edition,this text can also be used for self-study by technical professionals since it discusses engineering issues in algorithm design as well as the mathematical aspects. In its new edition,Introduction to Algorithms continues to provide a comprehensive introduction to the modern study of algorithms. The revision has been updated to reflect changes in the years since the book's original publication. New chapters on the role of algorithms in computing and on probabilistic analysis and randomized algorithms have been included. Sections throughout the book have been rewritten for increased clarity,and material has been added wherever a fuller explanation has seemed useful or new information warrants expanded coverage. As in the classic first edition,this new edition of Introduction to Algorithms presents a rich variety of algorithms and covers them in considerable depth while making their design and analysis accessible to all levels of readers. Further,the algorithms are presented in pseudocode to make the book easily accessible to students from all programming language backgrounds. Each chapter presents an algorithm,a design technique,an application area,or a related topic. The chapters are not dependent on one another,so the instructor can organize his or her use of the book in the way that best suits the course's needs. Additionally,the new edition offers a 25% increase over the first edition in the number of problems,giving the book 155 problems and over 900 exercises thatreinforcethe concepts the students are learning.

21,651 citations

Journal ArticleDOI
TL;DR: An important feature of the method is that arbitrary adaptation data can be used—no special enrolment sentences are needed and that as more data is used the adaptation performance improves.

2,504 citations

Journal ArticleDOI
TL;DR: A framework for maximum a posteriori (MAP) estimation of hidden Markov models (HMM) is presented, and Bayesian learning is shown to serve as a unified approach for a wide range of speech recognition applications.
Abstract: In this paper, a framework for maximum a posteriori (MAP) estimation of hidden Markov models (HMM) is presented. Three key issues of MAP estimation, namely, the choice of prior distribution family, the specification of the parameters of prior densities, and the evaluation of the MAP estimates, are addressed. Using HMM's with Gaussian mixture state observation densities as an example, it is assumed that the prior densities for the HMM parameters can be adequately represented as a product of Dirichlet and normal-Wishart densities. The classical maximum likelihood estimation algorithms, namely, the forward-backward algorithm and the segmental k-means algorithm, are expanded, and MAP estimation formulas are developed. Prior density estimation issues are discussed for two classes of applications/spl minus/parameter smoothing and model adaptation/spl minus/and some experimental results are given illustrating the practical interest of this approach. Because of its adaptive nature, Bayesian learning is shown to serve as a unified approach for a wide range of speech recognition applications. >

2,430 citations

Journal ArticleDOI
TL;DR: A model for the transit assignment problem with a fixed set of transit lines is described, formulated as a linear programming problem of a size that increases linearly with the network size that solves the latter problem in polynomial time.
Abstract: We describe a model for the transit assignment problem with a fixed set of transit lines The traveler chooses the strategy that allows him or her to reach his or her destination at minimum expected cost First we consider the case in which no congestion effects occur For the special case in which the waiting time at a stop depends only on the combined frequency, the problem is formulated as a linear programming problem of a size that increases linearly with the network size A label-setting algorithm is developed that solves the latter problem in polynomial time Nonlinear cost extensions of the model are considered as well

753 citations

Frequently Asked Questions (2)
Q1. What are the contributions in "Passenger route guidance system for multi-modal transit networks" ?

This study develops a multi-modal passenger route guidance system, called eFinder, to aid travelers with their combined mode-route choices. The authors discuss the architecture and features of this system in this study. This system forms a platform for disseminating public transit information and should complement further development and use of the public transport network by enabling travelers to make informed choices. 

Indeed, as a new prototype, there are several directions wherein further research and development are beneficial. Such models, once calibrated, can be used to study pricing and operation strategies for individual companies or the entire multi-modal network as a whole. The short message service ( SMS ) or multimedia message service ( MMS ) can be useful in this regard. Which part of the calculation can be performed offline and stored, and which part must be performed in real-time need to be carefully allocated.