Content uploaded by Avishai Ceder
Author content
All content in this area was uploaded by Avishai Ceder on Sep 30, 2016
Content may be subject to copyright.
Smart Feeder/Shuttle Bus Service
19
Smart Feeder/Shuttle Bus Service:
Consumer Research and Design
Y. B. Yim, University of California at Berkeley
Avishai (Avi) Ceder, Technion-Israel Institute of Technology
Abstract
While long-haul express transit is gaining ridership, consumers are increasingly expe-
riencing limited access to express transit due to saturated parking at and around
stations. e smart shuttle concept was introduced to provide easy access to express
transit. Smart shuttles will be equipped with advanced public transit system tech-
nologies to track shuttle vehicle locations and disseminate up-to-the-minute shuttle
arrival information to consumers. e first step toward deployment of the smart
shuttle service was a market study of short-haul feeders. is article presents the
results of a telephone survey of randomly-generated Castro Valley households. Castro
Valley is a suburban community in the San Francisco Bay Area, and many residents
commute by BART (Bay Area Rapid Transit). e survey suggested that three fifths of
the survey participants were likely to take the shuttle to the BART station. e interest
in using the smart shuttle service is strongly associated with gender, auto ownership,
ethnicity, and employment status. Females were more interested in taking a shuttle
than males. Employed people were more likely to use BART because of the shuttle.
Households without a car or with fewer cars were more interested in taking a shuttle.
e cost, travel time, and reliability of the service are the most important attributes
in the design of a shuttle. e subsequent phase of this research will be a field test
of the smart shuttle with optimal routing solutions. e value of the research is the
evaluation of the field test, which will assess the improvement of BART access and
the cost-effectiveness of the short-haul feeder operation. Ideally, this smart feeder/
Journal of Public Transportation, Vol. 9, No. 1, 2006
20
shuttle system will provide advanced and attractive service that operates reliably
and relatively rapidly and acts as part of the passenger door-to-door chain with
smooth and synchronized transfers. In order to approach the design of this innovative
feeder/shuttle system, new integration and routing concepts are presented based on
the consumer research.
Introduction
A growing concern for public transportation is its inability to encourage people
to switch their mode of transportation from solo driving to shared driving. As cit-
ies expand, transit ridership decreases while auto ownership increases. Although
overall transit ridership is declining in cities, an encouraging trend is increased
ridership in long-haul express bus or rail transit. When long-haul express transit
systems were built in the 1970s and 1980s in California, parking facilities were
also provided for riders to park their cars and ride a train. e concept of “park
and ride” was readily accepted by the public, and a large number of commuters
preferred to take an express bus or train to avoid rush-hour traffic and prohibitive
parking costs. As regional economies grew and more jobs became available, com-
muters increasingly relied on the express transit service (e.g., in the San Francisco
Bay Area). In 1999, the Bay Area Rapid Transit (BART) carried 285,000 commut-
ers each weekday, compared to 255,000 in 1992 (11.7% growth over seven years).
Automobiles continued to be the major access mode to BART. Approximately
80 percent of the park-and-ride BART customers parked in the BART parking lot
while the remaining 20 percent were parked off-site around the BART station on
residential streets. As the BART parking lots became full as early as 7:00 A.M., the
overflow vehicles took up space on residential streets, inviting an increasing num-
ber of neighborhood complaints around the BART stations.
Some BART riders claimed that they have had to switch back to driving because of
the severe access problems with BART. All 39 BART stations have access problems.
Although local buses are able to serve BART stations at certain times, they do not
necessarily meet the needs of BART customers’ schedules. Local bus and BART
schedules are not well synchronized and, thus, transfer times are often unneces-
sarily long. Long waits are one of the major reasons people do not want to travel
by public transit.
A smart shuttle/bus is an alternative travel method to personal vehicle or bus
transit for short-haul feeders. e smart shuttle concept has the potential for
Smart Feeder/Shuttle Bus Service
21
improvement of transportation accessibility for those who may not want to drive
or take a bus to express transit stations. Features of the smart shuttle will include
both fixed- and flexible-demand responsive systems based on the time and loca-
tion of the service, use of Advanced Public Transit Systems (APTS) technologies
for timely dissemination of the shuttle arrival time, and economical operation of
the smart shuttle system. Smart shuttle vehicles will be tracked via automated
vehicle identification (AVI) system and Global Positioning System (GPS). Ide-
ally, smart shuttles would provide attractive feeder and distributor services with
advanced transportation technologies contributing to transit operations that are
reliable, productive, and efficient. Smart shuttles could possibly provide door-to-
door service with smooth and seamless operation and synchronized scheduling
between long- and short-haul transit operations.
e short-haul feeder study as a whole has four broad objectives:
1. latent demand study (or market study) of short-haul feeders,
2. design of innovative routing strategies,
3. deployment of a smart shuttle for a field test, and
4. evaluation of the smart shuttle system.
e latent demand study is being performed using the survey research method.
e routing strategy study was done with simulation. e field test and evaluation
will be done following the first two studies.
is article presents the findings of the latent demand study, which is concerned
with consumer response to a smart shuttle system for short-haul feeders. e
article provides an understanding of the user side of short-haul feeder service. e
study investigated service attributes that would attract consumers. e decision
to take transit will depend on trade-offs between personalized transit and the
personal vehicle. If the goal of a demand responsive transit service is to capture
the driver population, a personalized transit service needs to be as convenient as a
personal vehicle. If the goal is to improve the transit service for those who are cap-
tive transit users, personalized transit needs to be better than fixed-route service.
e central issue is what attributes of personalized transit will attract consumers
and what segment of the population will use it. e target population of potential
users could include commuters, the elderly, persons with disabilities, and chil-
dren.
Journal of Public Transportation, Vol. 9, No. 1, 2006
22
e present study investigated desired attributes that will attract consumers. Key
research questions were:
1. What kind of service features will likely attract consumers enough for them
to switch from their single-occupancy vehicle (SOV) to a smart shuttle?
2. Who will use it and why?
e objectives of the latent demand study are:
• to assess the travel characteristics and profiles of potential users of the smart
shuttle with respect to socioeconomic variables, and
• to identify attributes necessary to support the smart shuttle system, includ-
ing acceptable waiting time, number of stops, size of shuttle vehicles, travel
time, and fare structure.
e findings of the demand study were considered in the design of the Castro Val-
ley’ smart shuttle system to improve BART station accessibility with new integra-
tion and routing concepts.
Castro Valley, a community within the incorporated County of Alameda in Cali-
fornia was selected for a field test. e reasons for selecting Castro Valley were:
• e community is one of the fastest growing suburban communities in
northern California.
• BART and Alameda Contra Costa Transit (AC Transit) serve Castro Valley.
• e community experiences severe access problems to the Castro Valley
BART station.
• Castro Valley has a severe shortage of parking around the BART and AC
express transit stations.
• There is a growing concern with overflow parking on neighborhood
streets.
While the innovative routing strategies are explained in this work, the simulation
study for these strategies for the Castro Valley community is presented by Ceder
and Yim (2002).
Background
An important issue regarding a smart shuttle is its design based on a good under-
standing of user needs and desires and how new technologies can enhance the
Smart Feeder/Shuttle Bus Service
23
smart shuttle service. In the past, dial-a-ride or door-to-door paratransit played
a vital role in North American transit planning in providing equitable transporta-
tion services to elderly and handicapped persons. But these concepts do not offer
transit services accessible to the entire population that needs mobility (Borndorfer
et al.1999; Ioachim et al.1995). Two user types were identified by Melucelli et al.
(2001), “passive users” and “active users.” Passive users make use of traditional
transit (i.e., boarding and alighting at compulsory stops). No reservation is neces-
sary since vehicles are guaranteed to serve each compulsory stop within a given
time window. Active users ask for a ride while boarding or alighting at an optional
stop. Active users must issue a service request and specify pick-up and drop-off
stops as well as earliest departure and latest arrival times. In this study, transit
vehicles have to be rerouted and scheduled to satisfy as many requests as possible,
complying with passage-time constraints at compulsory stops, while between two
compulsory stops optional stops can be activated on demand. e method used
in this study integrates mathematical programming tools into a search framework,
taking advantage of the particular structure of the problem formulation.
Dial-a-ride problems customarily use classical vehicle routing heuristics as
described in Laporte (1992), Shen et al. (1995), Savelsbergh and Sol (1995), and
Cordeau et al. (2000). ese methods are rooted in arc and node manipulation,
which generally is based on insertion, deletion, and exchange of stops in and out
of a current tour. e computation of an upper bound in finding the optimal dial-
a-ride solution is not a trivial issue. e linear relaxation of any arc-based integer
linear programming model provides, to some extent, a loose bound. erefore,
heuristics are necessary to cope with practical routing problems.
e findings of the demand study presented in this article are used to construct
simulation models for the development of routing strategies and generation of
optimal solutions to the smart shuttle services problem. A few studies make use
of simulation as a tool to devise satisfactory routing and scheduling solutions.
Two types of simulation studies can be traced in the literature. e first type
is the research conducted by Wilson et al. (1970, 1971) for evaluating various
heuristic routing rules and algorithms used in a computer-aided routing system.
ese studies were developed for mainframe computers and have limitations in
handling large-size road networks with different routing strategies. e second
research type, by Fu (2001) and his team, considered the use of advanced technol-
ogies. eir studies use a simulation model, Sim-Paratransit, which was developed
to evaluate advanced paratransit systems with AVL (automatic vehicle location)
Journal of Public Transportation, Vol. 9, No. 1, 2006
24
and CAD (computer-aided dispatch) systems. e ability to track continuously
a transit vehicle’s location enables the use of intelligent paratransit systems that
contribute to the operation of the paratransit systems at a significantly improved
level of productivity and reliability (Fu 2001).
Other advanced technologies include Advanced Traveler Information Systems
(ATIS). ATIS is aimed at providing timely and accurate traveler information
and bus or shuttle arrival. Although nearly a decade of ATIS research has been
undertaken, very few projects have investigated the potential application of ATIS
to fixed-route transit services. In the past, most federally-sponsored ATIS Field
Operational Tests or Model Deployment Initiatives were directed at assisting driv-
ers in ways of changing their travel behavior with advanced traveler information.
e potential application of ATIS technologies to the bus transit service is signifi-
cant, yet ATIS research in the transit area has lagged behind driver-oriented ATIS
studies. Many transportation policy-makers and practitioners believe that transit
users can benefit from getting real-time bus or shuttle information. e present
study uses ATIS technologies for the Internet dispatch of up-to-the-minute smart
shuttle arrival time so that end-users can readily retrieve shuttle information via
the Internet. Synchronized BART and shuttle arrival times will also be posted via
variable message signs at BART stations and other selected locations along shuttle
routes. Low-tech technologies also will be used. Cellular phones will be used for
the flexible-route flexible-schedule strategies. A customer will call the driver with
his or her location, then the driver deviates from the standard route schedule and
picks up the customer.
One of the reasons that people are reluctant to take transit is the uncertainty
associated with bus arrival times. ATIS can disseminate real-time bus schedule
information to those who are regular transit users as well as to the occasional tran-
sit rider. ATIS bus schedules also can attract those who have seldom or never used
transit. When the risk of taking transit is reduced to an acceptable level, people will
shift from driving to riding bus transit. Moreover, reasonably accurate arrival times
or travel times would increase the confidence level of bus transit operators.
Like fixed-transit service, demand responsive transit service suffers from low rider-
ship and high operating costs. As was evidenced in most demonstration projects,
demand responsive transit service was not cost-effective, with two factors contrib-
uting to this. From the operator perspective, either the design concept had flaws
or appropriate computer technologies were not available to efficiently operate
the system. From the user perspective, operators did not consider the needs of
Smart Feeder/Shuttle Bus Service
25
the users. Most of these systems provided the services without asking what type
of demand responsive service would attract consumers. While recognizing the
importance of operating system efficiency, advanced routing algorithms or com-
puting technologies may not necessarily generate increased ridership.
Previous field tests of demand responsive transit services showed their inability to
draw enough consumers to support the system even with public subsidies. Dial-a-
ride in the Boston area, a service that was later tested in other urban regions, had
similar problems of low ridership and high operating cost (Multisystems 1977;
Dave Systems 1977; Huron River Group, Missouri Transportation Associates,
Bishop Engineers 1977; Maine Department of Transportation 1986; TransVision
Consultants 1993). e reason was simply that the system was not attractive to
consumers. For consumers to be attracted to the new dial-a-ride service, it had to
be better than what they were using (i.e., light rail, bus transit, personal vehicle)
(Urban and Hauser 1993). Yet there is little understanding of consumer behavior
regarding what will make people favor demand responsive transit over other
modes of transportation.
With the advance of computer and communications technologies, it is now pos-
sible to improve the demand responsive transit system with up-to-the-minute
bus/shuttle arrival time and seamless operations between a short-haul feeder and
long-haul express systems.
Methodology
Studies on why people do or do not use transit are numerous. Some people may
have no option but to take transit (captured riders without a car), and others may
take transit because the cost of parking is prohibitive or taking transit is more
convenient than driving. We know a great deal about who the BART customers
are and who the AC Transit customers are and why they use these services; how-
ever, we do not know why people would use smart shuttles. Since smart shuttles
currently do not exist in the Castro Valley community, our objective is to discover
consumer reaction to this new proposed service.
To obtain the needed consumer information, the study used the survey research
method. e test market was identified as being within a 2-mile radius of the Cas-
tro Valley BART station. Four hundred telephone interviews were completed in
this market area using a random-digit-dial sample based on the 1990 census tract
information. (e 2000 census tract information was not available at the time of
Journal of Public Transportation, Vol. 9, No. 1, 2006
26
this market research.) During the last two weeks of September 2001, telephone
interviews were conducted using the Computer Aided Telephone Interview
(CATI) technique. e criteria used for screening survey participants were that the
potential participants had to:
• be 18 years old or older,
• be a permanent resident of the house called,
• say BART was a possible means of transportation for them, and
• commute or make their most frequent trip away from home by some means
other than walking or bicycling.
e margin of error for a 400-respondent sample is + 5.0 percent at the 95 percent
level of confidence.
e survey questions included the following topics:
• trip characteristics,
• mode of access transportation to BART,
• willingness to use a smart shuttle,
• willingness to pay for the service,
• desired attributes in the shuttle service, and
• demographic characteristics of survey respondents.
e study identified features that would attract consumers in terms of the routing
characteristics (i.e., intermediate stop options, express service), travel time, wait-
ing time, number of stops, and willingness to pay for the shuttle service.
Survey Results
e results of the survey are presented in several parts. ese include (1) demo-
graphic characteristics, (2) willingness to use the smart shuttle service, (3) will-
ingness to pay for the service, and (4) attributes that would enhance the shuttle
service.
Demographically, the sample was predominantly white, highly educated, and
financially well off. e sample is a representation of those who live in the sub-
urbs of the San Francisco Bay Area. is part of Alameda County is among the
middle- and upper-middle income communities in the Bay Area and commuters
Smart Feeder/Shuttle Bus Service
27
typically drive to well-paying professional or highly-specialized jobs in Oakland or
San Francisco.
Most people have the flexibility to choose a transportation mode to get to work,
school, or other destinations. However, some people are constrained by their jobs,
schools, or other reasons, making it infeasible to consider certain travel modes.
Survey participants were asked whether they had transportation options using
public transportation. When the respondent did not have the option of taking
BART, the survey was terminated. Surveying those with an option to take BART
was important because the research interest was to assess the market for a shuttle
service.
Of the participants, 64 percent commuted to work, 7 percent to school, and 2 per-
cent both equally. More than 57 percent of the participants have flexibility in the
time they start work. More than 80 percent start their work before 9 A.M. while
less than 20 percent start work after 10 A.M. Among commuters, nearly four fifths
(78%) of the respondents drove to work alone while only 8 percent carpooled and
19 percent took public transit. Ten percent of the public transit users were those
who parked their cars and rode BART.
Travel characteristics of commuters were significantly different from noncom-
muters (p<.05). Among noncommuters, more people tended to drive alone than
commuters (85%), and few people carpooled (8%), took public transit (5%), and
used the park-and-ride facility (2%). Two thirds (37%) of the noncommute trips
were made for shopping, 10 percent were for driving children, 10 percent were
for attending medical or dental appointments, and 18 percent were for social or
recreational events. e remaining 19 percent were for personal business.
Most (72%) of those who commuted with BART drove to the BART station by
personal vehicle. Only 3 percent carpooled and 17 percent walked to the BART
station. e median commute time by BART was 38 minutes, and the median
travel time to get to the BART station was 10 minutes. e mean cost of public
transit per day was $4.50; per week, $31.23; and per month, $71.82.
More than 80 percent of respondents said that parking was free for them; 13 per-
cent said they personally pay for parking; 5 percent reported that their company
pays for parking; and 2 percent noted both they and their company pay their
parking. However, there was no strong association between the parking situation
and the mode of transportation, at least in this survey. e survey also showed that
the cost of parking was relatively lower than the parking price in the central part
Journal of Public Transportation, Vol. 9, No. 1, 2006
28
of San Francisco or Oakland. e mean parking cost per day was $8.11; per week,
$32.50; and per month, $133.16.
One of the critical problems for BART operation is access to BART. Recently, some
residents of Castro Valley have said that they would consider using BART for their
commute to work or school but the parking lot at the station always seems full,
and it is difficult to find any other suitable parking near enough to the station. e
question was how many people do not take BART because of the parking situa-
tion. Of those commuters who do not use BART, 16 percent said that they do not
take BART because of the parking situation; 84 percent said they do not take BART
for some other reason.
Demand for a Shuttle Service
After a brief description of a BART shuttle service was read to participants, they
were asked to respond to the question how interested were they in using the
shuttle service. Afterwards, they were asked about the cost of the service wait
times, van size options, number of stops for pick-ups and drop-offs—all of which
are important attributes for designing a shuttle. Participants were then asked how
likely was it that they would use the service, given that the cost of the shuttle, wait
time, trip time, and scheduling were acceptable.
Interested in Using a Shuttle
e survey asked the following question about a neighborhood shuttle service to
provide easy access to BART: “Suppose a shuttle service were available that pro-
vided round-trip transportation to the closest BART station from a pick-up loca-
tion near your home. e service would use comfortable, air-conditioned vans,
and pick-ups would be scheduled for convenient times throughout the day and
would be coordinated with BART train schedules. How interested would you be
in this type of shuttle service, without considering the cost?” Using a 1-to-5 scale
where “1” meant “not at all interested” and “5” meant “very interested,” one quar-
ter of the respondents said that they are very interested in using the shuttle service
and one third said they are not at all interested in using the service. Approximately
half of the respondents stated that they are interested in using the shuttle service
to the BART station (Table 1).
Among the reasons for no interest in using the shuttle were:
Smart Feeder/Shuttle Bus Service
29
• e BART station is close enough to walk (24%)
• Need a car for work and errands (25%)
• BART is not convenient (38%)
Interestingly, more than 70 percent of the BART users travel at least 1 mile to get
to a BART station, while only 27 percent live within a six-block radius of a BART
station. ere was no association between the distance from the BART station and
driving to the BART station. People living three or four blocks from the BART sta-
tion drove to the station. People living miles from the BART station mostly drove
to the BART station. Very few took public transit or carpooled.
Approximately 40 percent of the participants expressed a high likelihood of using
the shuttle service. ey also said they would be more likely to use BART if a
shuttle service were available. When asked how many days a week participants
thought they would use BART because of the shuttle service, they responded that
they would use BART eight or more times a month.
Attributes
Several attributes were investigated with respect to the design of the shuttle ser-
vice. Among them were the number of pick-ups and drop-offs, the size of a shuttle
vehicle, acceptable number of riders, travel time, and wait time.
Important Attributes for Shuttle Design. For designing a shuttle service, we asked
respondents to name the three most important attributes in order of priority. Par-
ticipants said that the most important attribute was the cost of the shuttle service
(mean 2.29). e second most important was overall travel time, including the
waiting time for the shuttle either at BART or the pick-up location (mean 2.29).
Table 1. Interested in Using a Shuttle
5 Very interested 24.9
4 Somewhat interested 12.3 53.1
3 May be interested 15.9
2 Somewhat uninterested 13.9 46.9
1 Not at all interested 33.0
Journal of Public Transportation, Vol. 9, No. 1, 2006
30
e third most important was the on-time reliability of the service at the pick-up
location or at the BART station (mean 1.61).
Pick-ups and Drop-offs. When asked the maximum number of pick-ups that
should be allowed per trip to the BART station, most people expected four to five
pick-ups (median five pick-ups) on the way to the BART station. Similarly, they
expected four to five drop-offs on the way home (median five drop-offs).
Maximum Number of Riders in a Van. When asked the maximum number of peo-
ple each van should hold, respondents said that approximately 10 riders (median)
on each trip would be desirable.
Travel Time. When asked if the average travel time to the BART station were
slower than it currently is, would they still use the BART shuttle service? nearly one
third (29%) said they would take the shuttle, while half (53%) said they would take
the shuttle if it takes about the same time. Only 12 percent said they would take it
if it were faster. When asked if the average travel time to the BART station were 20
minutes longer than it takes currently to get to the BART station, would they use
the shuttle to the BART station? one fifth (20%) responded that they would use
it if it took 20 minutes longer, 20 percent said they would if 15 minutes longer, 40
percent said they would if 10 minutes longer, 20 percent responded they would
if 5 minutes longer. e survey suggests that people are willing to accept a longer
travel time using a shuttle for whatever the benefits they perceive.
Arrival Time and Schedule Information. One of the reasons that people are hesi-
tant to take transit is the uncertainty associated with bus arrival times. ATIS can
disseminate real-time bus schedule information to those who are regular transit
users as well as to the occasional transit rider. ATIS bus schedules also can attract
those who seldom or never used transit in the past. When the risk of taking transit
is reduced to an acceptable level, people will shift travel mode from driving to bus
transit. Reasonably accurate arrival times or travel times would increase the confi-
dence level of bus transit operators. A strong relationship between transit agencies
and ATIS could help Metropolitan Planning Organizations (MPOs) to improve
the overall Bay Area transportation system. e MPO’s largest concern is how to
change mode choice decisions from SOVS to high-occupancy vehicles (HOVs).
Cost. e cost question for riding a shuttle was constructed to ask about the high-
est price first and then subsequently lower prices. For the question: “Suppose the
cost for the shuttle service were $1 per one-way trip, how likely would you be to
use this service? Would you say that you definitely would use the service, prob-
Smart Feeder/Shuttle Bus Service
31
ably would use this service, might or might not use the service, probably would
not use the service, or definitely would not use the service?” Responses showed
that approximately two thirds (61%) of the respondents would be interested in
taking a shuttle at the price of $2 for a one-way trip, and half said they would be
interested if the cost were $1.
As expected, consumer interest in using the shuttle service is highly elastic with
respect to the cost of the shuttle service (Table 2, Figure 1). However, it is found
that the price elasticity is not directly proportional to the cost of the shuttle ser-
vice. Willingness to use the service is significantly different between the low and
the high cost of the shuttle service (p<.05).
Table 2. Willingness to Pay for the Shuttle Service
(%) (%) (%) (%) (%) (%)
5 Definitely would 4.4 1.8 3.5 8.1 14.8 7.0
4 Probably would 12.5 10.6 21.6 32.9 36.4 27.9
3 Might/might not 19.4 18.6 24.6 30.9 33.0 41.9
2 Probably would not 32.6 39.4 28.1 16.1 11.4 14.0
1 Definitely would not 31.1 29.6 22.5 12.1 4.5 9.3
Statistical significance p<.05 p<.05 p<.05 p<.05 p<.05
Figure 1. Willingness to Pay for the Shuttle Service
Journal of Public Transportation, Vol. 9, No. 1, 2006
32
Frequency of Using the Shuttle Service. When asked how often they would be will-
ing to use the shuttle service if the cost were acceptable, most respondents said
that they would use the service two to three times a week (mean 2.57, median 2).
Wait Time. Questions about wait time were posed at the same time as the cost of
the shuttle service. It was asked if the waiting time is 20 minutes, 15 minutes, 10
minutes, and 5 minutes, how likely would the participants be to use the shuttle
service? e answers were precoded in five scale responses: “definitely would,”
“probably would,” “might/might not,” “probably would not,” and “definitely would
not.” e survey showed that the longer the wait time, the less willing people are
to take the shuttle, but there was not a significant difference between a 10- and a
5-minute wait time (Table 3, Figure 2). is suggests that half of the shuttle users
are willing to accept a 5- to 10-minute wait time.
Table 3. Willingness to Wait for the Shuttle Service
(%) (%) (%) (%)
5 Definitely would 8.9 3.4 11.2 10.0
4 Probably would 26.8 25.7 40.8 41.7
3 Might/might not 21.5 25.2 28.0 35.0
2 Probably would not 21.6 24.0 10.4 6.7
1 Definitely would not 21.2 21.7 9.6 6.7
Figure 2. Willingness to Wait for a Shuttle Ride
Smart Feeder/Shuttle Bus Service
33
Payment Method. Participants were asked “How would you prefer to pay for the
service if you were to use the shuttle service?” More than 55.4 percent were inter-
ested in paying for the service on a per-user basis. Only 11.2 percent responded
in favor of the weekly-fee basis and 31.2 percent said they would work with a
monthly subscription arrangement.
Preferred Means of Receiving Information about the Shuttle. When asked about
how they would prefer to receive information about the shuttle service, including
the shuttle schedule, cancellation, and new services, 62 percent of the respondents
said that they would like to receive the information from a pamphlet. Approxi-
mately one third (30.9%) would like to receive it through the Internet, and only
5.9 percent would like to retrieve it by telephone.
Providing transit information through the Internet is more cost-effective than
over the telephone because Internet technology is widely deployed and does not
require human operators. Automation for telephone information still requires
technical improvement. e survey showed that nearly three quarters (73%) of
the participants had Internet access at home and 54 percent at work. While only
58 percent had personal cellular phones, 81 percent had a personal computer at
home, and 59 percent had one at work.
Benefits of the Shuttle. Because the question “What would be the biggest ben-
efits personally of using the shuttle service?” was an open-ended question and
accepted up to three responses, the percentages shown in this section are not
mutually exclusive. A variety of personal benefits of the shuttle service were men-
tioned. Among the personal benefits were:
• convenience, including no need to park (25%), avoid walking in bad weather
(2%), avoid wear on vehicle (21%), and others (30%),
• safety, including reduced stress and anxiety (8%), less chance of an accident
(2%), avoidance of traffic fights (18%),
• travel time savings (14%),
• less cost (18%),
• reduced pollution (7%), and
• chance to meet people and socialize (2%)
e survey suggested that most people perceive benefits from the shuttle service.
It would be convenient for them and could save travel cost and time and increase
safety and reduce stress.
Journal of Public Transportation, Vol. 9, No. 1, 2006
34
Respondents also mentioned that there would be a great deal of social benefit to
be gained from the shuttle service. Among the societal benefits were:
• reduction of traffic congestion (52%),
• reduction of air pollution (40%),
• easier to get around, greater mobility, greater accessibility (29%),
• less crowded parking lots at the long-haul express transit station (10%),
• saving of money and lower taxes (7%), and
• reduction in accidents (8%)
The Likelihood of Using the Shuttle Service
After a series of questions about design attributes were posed, including the
acceptable fare, wait time, the number of pick-ups and drop-offs, and the size of
a shuttle vehicle, participants were asked again: “If the shuttle service cost what
you are willing to pay and has acceptable wait times, trip length, and scheduling
times how likely do you think you would be to use the shuttle service to get to
and from the BART station?” Using the 5-point scale where “1” meant “not at all
likely” and “5” meant “very likely,” 57.2 percent said they would be likely to use the
shuttle service and 29.2 percent said they may use it. Only 13.7 percent said they
were not likely to use the service. is response is significantly different (P < .05)
from the earlier question concerning how interested respondents were in using
the shuttle service without considering the cost. e latter responses, after learn-
ing about the shuttle attributes, were far more receptive to the shuttle service
than the former responses (Table 4). In the latter case, nearly 60 percent of the
respondents said that they would be likely to use the shuttle, while in the former
case the same sample indicated that less than 40 percent would be interested in
using the shuttle.
In response to the interest in using the shuttle, there was no difference between
commuters and noncommuters. Similarly, there was no difference between com-
muters and noncommuters in their likelihood of using the shuttle after learning
more about its design.
When asked whether they think they would be more likely to use BART because
of the shuttle service or whether it would make no difference in how frequently
they use BART, 43.3 percent of those surveyed said that they would be more likely
to use BART if a shuttle is provided, 23.3 percent said much more likely, 17 percent
Smart Feeder/Shuttle Bus Service
35
said somewhat more likely, and 56.8 percent said that the shuttle would not neces-
sarily cause them to take BART more often.
e analysis showed that public interest in taking a shuttle service was closely
associated with gender, automobile ownership, ethnicity, and employment status.
Females were more interested in taking a shuttle than males (P < .05). Employed
people were more likely to use BART because of the shuttle service (P<.05). House-
holds without a car or with fewer cars were more interested in taking a shuttle (P
< .05). People among the Asian-American, Hispanic, Native American, and mixed
race groups were marginally more interested in taking transit than the white or
black race (p=.058). However, further analyses showed that the likelihood of taking
a shuttle was not closely associated with gender, automobile ownership, ethnic-
ity, and employment status. ere was no difference between commuters and
noncommuters in their interest or the likelihood of taking a shuttle for the BART
service. Nonetheless, the study found that availability of parking at their workplace
was closely associated with taking BART (P < 0.05).
Smart Feeder/Shuttle Design: Routing Strageties
Once the major elements of the smart feeder/shuttle transit service are defined,
attention should be given to smart routing strategies. ese strategies represent
the flexibility and, to some extent, part of the attractiveness of the transit system.
Ten routing strategies were investigated in this work:
1. fixed route with a fixed schedule (timetable) and fixed direction;
2. fixed route with a flexible (demand-driven) schedule, fixed direction;
Table 4. Comparative Response (P < .05)
%) (%)
5 Very interested 24.9 30.3%
4 Somewhat interested 12.3 26.9
3 May be interested 15.9 29.2
2 Somewhat uninterested 13.9 12.2
1 Not at all interested 330 1.4
Journal of Public Transportation, Vol. 9, No. 1, 2006
36
3. fixed route with a flexible schedule, bidirectional;
4. fixed route, flexible schedule, fixed direction, with a possible short-turn;
5. fixed route, flexible schedule, bidirectional, with a possible short-turn;
6. fixed route, flexible schedule, fixed direction, with a possible shortcut;
7. fixed route, flexible schedule, bidirectional, with a possible shortcut;
8. fixed route, flexible schedule, fixed direction, with possible short-turn and
shortcut;
9. fixed route, flexible schedule, bidirectional, with possible short-turn and
shortcut; and
10. flexible (demand responsive) route with a flexible schedule.
Fixed direction means that the shuttle will always maintain the same direction of
travel (same sequence of stops), whereas bidirectional allows for having the flex-
ibility to select the direction based on real-time demand information. e term
“shortcut” means that, based on certain loading threshold and synchronization
criteria, the shuttle will not continue its fixed route and, instead, will use the short-
est path (minimum travel time) to arrive at the train station. e loading threshold
is a given (input) number of passengers on board the shuttle. e synchronization
criterion means matching the shuttle’s new (shortcut) arrival time with an earlier
train than that originally planned if the entire route is completed. e term “short-
turn” means that based on certain loading threshold and synchronization criteria,
the shuttle will not continue on its fixed route. Instead, it will turn around and
arrive at the train station in the opposite direction, with the possibility of picking
up passengers who were too late to be picked up when the shuttle passed through
the station previously. e loading threshold and synchronization criteria for the
short-turn strategy (including the consideration of more pick-ups) are the same
as for the shortcut strategy. Each strategy allows the flexibility of the other; that is,
the loading threshold of the shortcut strategy is higher than the loading threshold
of the short-turn strategy. If the latter is reached and there is the possibility of
picking up x passengers (after turning around), where x is equal to or greater than
the difference between the two loading thresholds, then the short-turn strategy
is recommended.
Figure 3 depicts the 10 strategies on a small network with two shuttle routes, one
with a dashed line and one with a dotted line. e clock on the upper-right-hand
side exhibits the fixed schedule (in only one strategy); when crossed with x, it
means a flexible schedule situation. Arrows in both directions of the route means
Smart Feeder/Shuttle Bus Service
37
Figure 3. Routing Strategies Considered on a Small Network Example
Journal of Public Transportation, Vol. 9, No. 1, 2006
38
a bidirectional situation. It can be seen in Figure 3 that the lines with the arrows
deviate from the fixed route in the shortcut strategy. e arrows turn around at a
certain point of the network in the short-turn strategy, while both representations
appear in the strategy involving a possible combination of shortcut and short-turn
runs,. e last strategy is for a DRT-type of service, allowing for the creation of a
new route every time, based on the trip bookings.
e idea of covering almost all possible practical routing strategies stemmed from
the need to arrive at user desires and understandings. Certainly, there is no inten-
tion that all strategies be used at the same time; rather, the idea was to examine
which strategy was best for a given demand pattern and magnitude while taking
into consideration the real-time traffic situation in the area of the shuttle’s trips.
A simulation model was devised for that purpose. is simulation tool, to be
explained in the next session, enables a comparison of the various strategies, based
on the following measures:
• sum of total time (in passenger-hours) from passenger pick-up to train-
departure times,
• sum of total time (in passenger-hours) riding the shuttle vehicle,
• sum of total waiting time (in passenger-hours) for the train,
• sum of total waiting time (in passenger-hours) for the shuttle vehicle, and
• total number of transit vehicles (by number of seats) required to meet the
demand.
ese measures of travel and waiting times and number of vehicles characterize
the effectiveness and efficiency of each strategy. Certainly, the strategy selected for
a given demand is the one with the minimum weighted travel and waiting times
(user perspective) and the minimum number of vehicles (operator perspective).
ese routing strategies underwent a simulation process explained and inter-
preted in Ceder and Yim (2002).
Once the analysis of the feeder/shuttle service is completed, we recommend the
next step should be a pilot study, the implementation of which can follow, for
example, the 12 steps shown in Figure 4. ese 12 steps of Figure 4 can serve as
a framework for a master plan of a pilot where each outcome of a previous step
becomes an additional input to the next step except for step 6.
e pilot master plan starts with a demand analysis by time of day and day of week
to find the origin-destination pattern and consumer oriented features. e second
Smart Feeder/Shuttle Bus Service
39
Figure 4. Overview of Feeder/Shuttle Pilot Master Plan
step is to design the fixed routing and stop system and the third-to determine the
base frequencies and timetables for each route. e fourth step is to determine
the number and size of the feeder/shuttle vehicles and to create the chains of
trips (vehicle schedules) which will serve the fifth step of constructing the crew
schedules.
Journal of Public Transportation, Vol. 9, No. 1, 2006
40
e pilot plan continues in step 6 with the establishment of effective information
channels and instruments (e.g., telephone center, internet, newspapers, radio, TV,
mail leaflets) which will lead to the development of a user-friendly communication
procedures between the users and the operator in the next step. Step 8 constructs
the DRT operational strategies without the use of the fixed routing/stop/schedule
system. Step 9 determines the testing scenarios of the pilot while step 10 presents
the process to select an adequate operator. Step 11 uses proper advertisement
tools to approach an operable pilot, and, finally, the last step of the plan aims at
improving the instruments, procedures and strategies with the use of innovative
ITS (Intelligent Transportation Systems) elements.
Conclusions
is article documented the survey results of Castro Valley, a suburban commu-
nity in the San Francisco Bay Area of California that is experiencing rapid growth of
housing developments along with the elderly and the young. ere was evidence
in the survey of demand for a smart shuttle service, as approximately half of the
commuters with the option to take BART would be likely to use such a smart
shuttle service. ere was no strong association between age or income group and
the likeliness of using the shuttle service. However, interest in using the service is
closely associated with gender, auto ownership, and ethnicity and employment
status.
e study suggests that there is a strong potential for the deployment of a smart
shuttle service in Castro Valley. e present study fills the gap in our understand-
ing of a potential market for a short-haul feeder system to support the long-haul
express transit. e Castro Valley study is our first field test in the San Francisco
Bay Area. e field test of the smart shuttle project is expected to be deployed in
the near future following interest among some of the Bay Area cities. e value of
the innovative transit service is in evaluation. e evaluation of the field test will
provide valuable insights into the technical validity of the smart shuttle and the
cost-effectiveness of the system.
In addition, this work attempts to construct a new idea for designing an inte-
grated smart feeder/shuttle bus service. Ideally, this smart bus system will provide
advanced and attractive feeder and distributor services that operate reliably and
relatively rapidly, and are part of the passenger door-to-door chain with smooth
and synchronized transfers. Ten different routing strategies are proposed with
Smart Feeder/Shuttle Bus Service
41
all the combinations of fixed/flexible routes, fixed/flexible schedules, one or
bidirectional concepts, and shortcut (shortest path) and/or short-turn (turn
around) concepts. Finally a 12-step implementation framework is shown to bridge
between the consumer research results and realization of the smart feeder/shuttle
bus design.
Acknowledgments
is is to acknowledge with deep grief that, after the completion of this paper, the
first author, Dr. Y. B. Yim, passed away. Her death was a shock and loss to all who
knew her personally and professionally. is paper is yet another remembrance
of Dr. Yim’s many achievements and contributions to the transportation com-
munity.
References
Borndorfer, R., M. Grotschel, F. Klostemeier., and C. Kuttner. 1999. Telebus Berlin:
Vehicle scheduling in a dial-a-ride system. In N.H.W. Wilson, ed., Computer-
Aided Transit scheduling, lectures notes in economics and mathematical sys-
tems 471. Berlin: Springer, 391–422.
Ceder, A., Y.B. Yim. 2002. Integrated Smart Feeder/Shuttle Bus Service, UCB PATH
Program publication.
Comsis Corporation (1988). Cost analysis methodology for demand responsive
service, Urban Mass Transit Administration and the Maryland Department of
Transportation.
Cordeau, J.-F., G. Laporte, and A. Mercier. 2000. A unified tabu search heuristic for
vehicle routing problems with time windows. Technical Report CRT-2000-03.
Montreal: Centre de Recherche sur les Transports, University of Montreal.
Dave Systems, Inc.1977. An integrated public transportation system for the cities
of Fremont, Newark, and Union City, the Tri-City Transit Implementation
Project.
Fu, L. 2001. Simulation model for evaluating intelligent paratransit systems. Trans-
portation Research Record 1760. Washington, DC: TRB National Research
Council, 93-99.
Journal of Public Transportation, Vol. 9, No. 1, 2006
42
Huron River Group, Missouri Transportation Associates, Bishop Engineers. 1977.
Dial-a-ride study: Report No.1, Dial-a-ride service potential in the Kansas City
Metropolitan Region.
Ioachim, I., J Derosiers, Y. Dumas, M. Solomon, and D. Villeneuve. 1995. A request
clustering algorithm for door-to-door handicapped transportation. Transpor-
tation Science 29: 35–139.
Laporte, G. 1992. e travelling salesman problem: An overview of exact and
approximate algorithms. European Journal of Operational Research 59: 231–
247.
Maine Department of Transportation. 1986. An analysis and action strategy for
select surface passenger transportation services in Maine.
Melucelli, F., M. Nonato, T.G. Crainic, and F. Guertin. 2001. Adaptive memory
programming for a class of demand responsive transit systems. In Stefan Voss
and Joachim R. Danuna, eds., Computer-aided scheduling of public transport.
Lectures notes in economics and mathematical systems 505. Berlin: Springer,
253–273.
Multisystems, Inc. 1977. State of Wisconsin Urban Mass Transit Demonstration
Program: Merrill project, Final Report.
Savelsbergh, M.W.P., and M. Sol. 1995. e general pickup and delivery problem.
Transportation Science: 17–29.
Shen, Y., J.-Y. Potvin, J.-M. Rousseau, and S. Roy. 1995. A computer assistant for
vehicle dispatching with learning capabilities. Annals of Operations Research
61: 189–211.
TransVision Consultants Ltd. 1993. Evaluation of Handyline: Vancouver’s auto-
mated paratransit information system. BC Transit.
Urban, G., and J. Hauser. 1993. Design and marketing of new products, 2nd ed.
Englewood Cliffs, NJ: Prentice Hall.
Wilson, N.H.M., J.M. Sussman, B.T. Higonnet, and L.A. Goodman. 1970. Simulation
of a computer-aided routing system (CARS). Highway Research Record 318.
Washington, DC: HRB National Research Council, 66–76.
Wilson, N.H.M., J.M. Sussman, H. K. Wong, and B.T. Higonnet. 1971. Scheduling
algorithms for a dial-a-ride system. Cambridge: Massachusetts Institute of
Technology.
Smart Feeder/Shuttle Bus Service
43
About the Authors
obtained her B.Arch. (architecture/structural) at the University of
Washington, M. Arch. (architecture/urban design) at the University of Washing-
ton, M.C.P. (city planning) at the Harvard University, and Ph.D. (Transportation
Engineering) at the University of California at Berkeley. She was a member of:
Intelligent Transportation Society of America ATIS Committee, Transportation
Research Board Committee on User Information Systems, Western Regional Sci-
ence Association, and Transportation Research Board. Dr. Yim has written articles
and reports on numerous transportation-related topics, including traveler infor-
mation and travel behavior, demand responsive transit systems, evacuation plan-
ning and emergency service, and consumer research in transit.
(ceder@tx.technion.ac.il) obtained his B.Sc. (1971) at the
Technion (industrial and management engineering), and M.Sc. (1972) and Ph.D.
(1975) at the University of California at Berkeley (major in transportation with
emphasis on operations research and human factors). Dr. Ceder has served in
the Technion since 1975 and MIT-Boston as a Visiting Professor since 1981. He
is a Principal Investigator at the Technion and is or was involved with teaching
and research projects at the University of California at Berkeley, M.I.T, Hong
Kong University of Science and Technology, Hong Kong Polytechnic University,
University of Tokyo, University of Rome, and at the Graduate School of Business,
Sydney University, focusing on public transport operations and service planning,
operations research, intelligent transportation systems, traffic safety and control,
and logistics.
Dr. Ceder has written more than 100 articles in refereed journals, book chapters,
proceedings and has authored three books: Network eory and Selected Topics in
Dynamic Programming, System Analysis as an Introduction to Operations Research,
and Public Transport: Operations and Service Planning (Lecture Notes). In addition,
he served as chief scientist of the Israel Ministry of Transport, Israel delegate at the
Transport Program of the European Community, president of the Israel Associa-
tion of Transportation Research, and a member of various organizing committees
of international symposia and workshops.
