Transport Reviews

ISSN: 0144-1647 (Print) 1464-5327 (Online) Journal homepage: http://www.tandfonline.com/loi/ttrv20

Requirements from vehicle routing software:

perspectives from literature, developers and the
freight industry

Nicolas Rincon-Garcia, Ben J. Waterson & Tom J. Cherrett

To cite this article: Nicolas Rincon-Garcia, Ben J. Waterson & Tom J. Cherrett (2017):
Requirements from vehicle routing software: perspectives from literature, developers and the freight
industry, Transport Reviews, DOI: 10.1080/01441647.2017.1297869

To link to this article: http://dx.doi.org/10.1080/01441647.2017.1297869

Published online: 05 Mar 2017.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=ttrv20

Download by: [Pontificia Universidad Javeria] Date: 06 March 2017, At: 03:47
TRANSPORT REVIEWS, 2017
http://dx.doi.org/10.1080/01441647.2017.1297869

Requirements from vehicle routing software: perspectives

from literature, developers and the freight industry
Nicolas Rincon-Garciaa,b, Ben J. Watersona and Tom J. Cherretta
a
Transportation Research Group, University of Southampton, Southampton, UK; bIndustrial Engineering
Department, Pontificia Universidad Javeriana, Bogota, Colombia

ABSTRACT ARTICLE HISTORY

Routing and scheduling software is part of the information and Received 22 October 2015
technology systems available to support the transport industry, Accepted 9 February 2017
and uses complex algorithms along with geographical
KEYWORDS
representations of the road network to allow better planning of Vehicle routing software;
daily collection and delivery schedules. This paper reviews the vehicle routing problem
evolution of routing and scheduling software, the algorithms used
along with reported barriers to wider take-up and potential
industry-driven improvements that could be made. A survey of
transport companies in the U.K. was conducted in order to
validate and prioritise the software capabilities that require the
most development according to the new challenges that the
industry is facing. Responses suggested that companies required
improved route optimisation to tackle congestion based on time-
dependent data and models, and greater accuracy in the
representation of the road network. Not considering congestion
leads to the underestimation of travel times and the production
of inaccurate schedules. Literature shows that operational
research techniques are available to solve problems that represent
real-world conditions, but research into the relative merits of
using time-dependent models needs to be undertaken. Data
might be improved by cooperation between government and
private sector.

1. Introduction
Logistics and transportation are considered key activities contributing to business success
where technological innovation is seen as being vital for obtaining better use of the road
network (European-Commission, 2012). New challenges have arisen in the transport indus-
try such as road congestion and regulation imposed by authorities to tackle the negative
impacts of transport. This coupled to the ever demanding customer and the move to
same-day delivery has meant that logistics providers have had to continually improve their
level of service (Ehmke, Steinert, & Mattfeld, 2012). The continuous growth in e-commerce
and home delivery over the last 16 years has imposed additional constraints on the indus-
try (Visser, Nemoto, & Browne, 2014) where carriers are trying to meet ever more exacting

CONTACT Nicolas Rincon-Garcia nicolas.rincon@javeriana.edu.co Transportation Research Group, University of

Southampton, Southampton SO17 1BJ, UK; Industrial Engineering Department, Pontificia Universidad Javeriana, Bogota,
Colombia
© 2017 Informa UK Limited, trading as Taylor & Francis Group
2 N. RINCON-GARCIA ET AL.

customer requirements using more dynamic schedules which are at the mercy of inaccur-
ate link travel times (Eglese, Maden, & Slater, 2006; Ehmke et al., 2012). In the U.K., 10% of
the retail spend is through online shopping and it is estimated that 12% of first time deliv-
eries to customers fail as a result of the consignee not being present at the delivery
address (Visser et al., 2014). This requires re-delivery attempts to be made and modifi-
cations to pre-planned schedules which costs carriers time. At the same time, there is a
need to handle product returns and recyclate within the urban environment and improved
reverse logistics schedules could see a reduction in empty running and an increase in
overall vehicle utilisation (Cherrett et al., 2012).
As a result, vehicle scheduling is an important element in transport operations to
achieve efficiency, where effective planning has to address multiple elements according
to the restrictions that the business model requires. Although it is a complex activity
that requires human expertise, a number of information and technology (IT) systems
are available to support drivers and planners such as computerised vehicle routing and
scheduling (CVRS) systems and satellite navigation (Sat-Nav) systems (DFT, 2010).
Rochat and Semet (1994) illustrate the benefits of supporting schedules with CVRS in a
Swiss company producing pet food and flour, by using software it was possible to
produce routes, in a matter of minutes, that complied with driving time regulation,
weight and access restrictions while reducing travelled distance, something that is time
demanding and difficult to achieve manually.
Despite the many benefits reported by the industry and the scientific literature, adop-
tion rates of CVRS is low. A survey conducted with 335 companies in 2010 by the U.K.
Department for Transport found that the adoption rate for companies with less than 10
vehicles in their fleet was 11% and for companies with more than 10 vehicles, 17%. The
reasons for not adopting CVRS were not suitable (49%), too expensive (33%), not war-
ranted for the nature of the operation (9%) and too complex (2%) (DFT, 2010).
This paper reviews the evolution of CVRS from the perspective of the users’ experi-
ence, investigating the software capabilities along with the perceived barriers to its
future development. Special attention is paid to the vehicle routing problem (VRP)
and the set of models that support the decision-making processes, aided by a survey
of logistics providers operating in the U.K. Previous work has tended to focus on: (i)
identifying VRP models that require further research from the theoretical perspective,
(ii) surveys to identify industry adoption rates and (iii) surveys of software providers
to identify software capabilities and approaches. This research identifies CVRS require-
ments based on industry experience and user needs, rather than from the mere per-
spective of the software vendor.

2. Evolution of CVRS systems

“Decision Support Systems (DSS) are computer technologies that can be used to support
complex decision-making and problem solving” (Shim et al., 2002, p. 111). Since the 1970s,
major technological breakthroughs have offered innovative solutions for vehicle planners
and more recently, individual drivers. In this section, a review of CVRS capabilities and the
respective technological components that support them are presented.
The three main components of DSS are a database, a model and a user interface. Since
the early 1970s, the evolution of DSS has been influenced by data processing, the
 TRANSPORT REVIEWS 3

microcomputer and network communication developments (Shim et al., 2002). Support-

ing the routing decision might require an additional element to represent the network,
geographical information systems (GIS) allow the manipulation and visualisation of geo-
graphical data (road characteristics) such as streets, roads, intersections, velocities, conges-
tion times, street names, length and address ranges (Bozkaya, Yanik, & Balcisoy, 2010;
Ioannou, Kritikos, & Prastacos, 2002; Sadeghi-Niaraki, Varshosaz, Kim, & Jung, 2011;
Weigel & Cao, 1999). Among the reported benefits in the implementations of CVRS are
(Tarantilis & Kiranoudis, 2002):

. transport cost reduction;

. reduction in fuel consumption and environmental impact;
. improved customer service;
. effective strategic planning;
. less reliance on individual skills and
. tighter control of distribution.

Although CVRS has advanced considerably since its first inception, manual intervention
in route and schedule design is still vital to the operation in order to prioritise or relax
certain constraints using the real-world knowledge of the logistics planner to provide
the most efficient solutions (Cegarra, Gacias, & Lopez, 2012). Despite this, many technol-
ogies have aided the development of CVRS tools particularly GIS for enabling detailed
road characterisation and mobile technologies to allow the tracking of vehicles in real
time and the transfer of routing information emanating from CVRS systems to the
driver (Figure 1) (Jung, Lee, & Chun, 2006).
The evolution of CVRS is presented in Figure 2. Initial schedules were based on obtain-
ing a minimum distance/time route that satisfied all the customer service requests. In this

Figure 1. Flowchart of the planning/delivering process with re-routing. Source: Jung et al. (2006).
4 N. RINCON-GARCIA ET AL.

Figure 2. Development of CVRS. Source: authors.

way, Dantzig and Ramser (1959) proposed a theoretical model for determining the
minimum distance routes for petrol tankers travelling between customers, so that each
site was only visited once while satisfying all demands and not exceeding the vehicles
capacities.
The development and adoption of GIS by businesses and government in the late 90s
(Keenan, 1998), allowed the inclusion of detailed road characteristics into commercial
CVRS. In a survey conducted in 2010 to CVRS providers, 86% reported GIS capabilities,
for example, displaying routes and maps, editable routes, address geocoding and turn-
by-turn directions (where routes between locations are described by presenting the
road segments and the turns) (OR/MS-Today, 2010). Road characterisation offers the possi-
bility to include into the model legal restrictions for vehicles such as maximum weight or
size on specific roads and time-related access restrictions (e.g. access restriction to certain
zones and roads due to congestion such as the London Lorry Control Scheme (FTA, 2002)).
Additionally, calculating the shortest paths between customers taking into consideration
the road network geometry and restrictions becomes an automated task (Jung et al.,
2006).
Mobile technologies, defined as portable devices that encompass hardware, software
and communications (Jarvenpaa & Lang, 2005), that utilise GIS concepts and global posi-
tioning systems (GPS) with real-time communication have brought a range of products to
the transport industry. Satellite navigation systems offer drivers turn-by-turn directions
when negotiating unfamiliar roads and 44% of companies with more than 10 vehicles
in their fleet reported using it (DFT, 2010). In cab, communications and GPS provide
vehicle tracking and status of deliveries in real time in order to re-route vehicles if necess-
ary (e.g. courier pick-ups and deliveries, food distribution) (Montemanni, Gambardella,
Rizzoli, & Donati, 2005). Additionally, it provides control over the operation with customers
able to access the system and get information in real time about the status of the delivery
and the location of the vehicle.
 TRANSPORT REVIEWS 5

User interaction has also evolved with the implementation of these technologies
through the improved design of the interfaces and systems. Ploskas, Ioannis, Jason, and
Nikolaos (2015) presented a CVRS that makes use of a specially designed hardware that
allows multiple users to interact in order to create the best possible route where traditional
input devices (e.g. mouse and keyboard) are replaced by a touch display. In a similar way,
an operative prototype was implemented by Cherrett et al. (2015) at Oxfam in order to
enhance its supply chain visibility. Drivers could retrieve and share information with
other drivers, schedulers and shop managers through a smartphone app. Managers
valued the app as it contributed to better decision-making with the addition of more
real-time information.
The technologies previously mentioned have also made possible the collection and
communication of data depicting the status of the road network; concepts such as intel-
ligent transportation systems (ITS) and advanced traveller information systems are based
on obtaining, processing and communicating information in order to make a better use of
transportation resources (Crainic, Gendreau, & Potvin, 2009; Khattak, Targa, & Yim, 2004).
Changing road conditions such as speed due to congestion is an issue for transport com-
panies where schedules become unreliable if congestion is not considered when planning
(Haghani & Jung, 2005; Kok, Hans, & Schutten, 2012). In a survey of transport managers in
California, more than 80% considered congestion “serious” or “critically serious” due to
unreliable travel times, increased costs and driver frustration (Golob & Regan, 2001).
CVRS providers report the utilisation of traffic information for routing (OR/MS-Today,
2014), which provides more reliable schedules and not considering congestion might
lead to the underestimation of travel time and missed deliveries (Haghani & Jung, 2005;
Kok et al., 2012). Although congestion can be predicted to a certain extent, recurrent
delays due to it are responsible for up to 80% of the total delays in peak hours
(Skabardonis, Varaiya, & Petty, 2003), and mobile technologies can allow drivers to receive
road information in real time in order to avoid unpredicted congestion (Cohn, 2009).
An important element in CVRS is the ability to provide control over the operation,
where creation of reports with a statistical module is a common capability (Drexl, 2012).
A new challenge in the industry involving data storage is the driving time regulation,
where drivers of vehicles over 3.5 tons are subject to mandatory breaks and limited
driving hours, the main aim of this policy is improvement of road safety by reducing
fatigue and drowsiness (Jensen & Dahl, 2009). The Tachograph is used to record the activi-
ties of drivers, and can be paired to CVRS in order to provide driving times and smooth out
driver workloads (Paragon, 2009).
In the U.K., the main established CVRS providers are DiPS, LogiX, Optrack, Paragon,
Roadnet, Descartes and Truckstops (DFT, 2005) and five out of the seven are participants
of the software survey conducted by OR/MS-Today (2016). Software characteristics
(routing functions, features and type o fleet) are shown in Table 1. Turn-by-turn instruc-
tions are supported by the five vendors along with capabilities for handling driving
time regulation, driver competency and geographic conditions. Regarding inclusion of
congestion in routing, 4 of the vendors provide capabilities to utilise historical traffic
time from mobile devices with only three vendors supporting some type of re-routing
based on real-time traffic conditions (data from Inix, Google Maps and HERE) (OR/MS-
Today, 2016). Vendors offer support to different types of fleets, each type might have
special requirements, for example, long haul requires information about driving time
 6
N. RINCON-GARCIA ET AL.
Table 1. Software characteristics of main CVRS vendors in the U.K.
Routing functions Algorithm considerations Type of fleet
Real-time
traffic Historical
(reassign travel Source
stops/re- time for for real- Long
Number of sequence routing time Local haul
companies Same Assigns Turn-by- stops/ from travel Driver pick-up Less-
using day re- Daily Weekly individual turn route turn- mobile time Driver Geographic hours and Than- Long Service
Characteristics software routing routing routing drivers instructions by-turn) devices data skills restrictions rules delivery Load Haul Courier Buses fleets
Software Optrak 1–100 x x x X x x x x x x x x x
Paragon 101–500 x x x x x x x x x x x x x x x
Roadnet +1001 x x x x x x Inix x x x x x
Descartes +1001 x x x x x x x Google x x x x x x x x
Maps
TruckStops +1001 x x x x x x x HERE x x x x x x x x
Source: OR/MS-Today (2016).
 TRANSPORT REVIEWS 7

regulation with restrictions for the week, overnight parking and resting, petrol stations,
cost of petrol, tolls, etc. (Rowell, 2012).
The characteristics of the algorithms used by the different vendors are shown in Table 2.
Solution methods for commercial routing software should be capable of solving large
instances (Drexl, 2012), and most vendors offer unrestricted size capabilities with
Paragon providing routing solutions up to 20,000 customer requests across 3000 vehicles.
Sörensen, Sevaux, and Schittekat (2008) argue that most developers in the industry prefer
a large set of heuristics rather than sophisticated scientific approaches that fall into the
category of metaheuristics that might take too long to solve large instances and are
capable of avoiding miopic behaviour associated with a reduced number of neighbour-
hood structures (e.g. metaheuristics are commonly implemented with few neighbourhood
structures). Additionally, this approach is useful for solving different types of VRP variants
required by the different logistic configurations present in the industry in reasonable
execution time. Vendors are capable of solving problems involving 1000 customer
requests in matter of minutes or even seconds. However, there remains the issue of the
quality (accuracy) of the algorithm. In a test carried out by an academic group using differ-
ent CVRS providers, significant differences in quality of solutions were found, with up to
10% difference between the best and worst schedules in instances involving only 100 cus-
tomer requests and difference increases with larger instances (Bräysy & Hasle, 2014;
Hallamaki et al., 2007). Relation between accuracy, heuristics and metaheuristics will be
reviewed in Section 4.2.
Although it seems that from the vendor perspective, there are a range of capabilities to
support the transport industry, there remain questions about the components that require
further improvements from the user’s perspective. The following section introduces a
survey conducted with companies whose freight operations made use of CVRS in order
to understand the requirements of the industry.

3. Survey
The survey was designed to be answered by planners and transport managers in U.K. com-
panies with freight transport operations that have implemented CVRS using an online
structured questionnaire. Companies were identified by business social networks, business
directories and the database for registered heavy goods vehicles. Nine hundred compa-
nies were contacted initially by phone in order to request information regarding charac-
teristics of their transport operations and willingness to participate in the study. It is
important to highlight that only companies already utilising CVRS could participate in

Table 2. Algorithm characteristics of main CVRS vendors in the U.K.

Optrak Paragon Roadnet Descartes TruckStops
Maximum size of NS: unlimited NS: 20,000 NS: unlimited NS: unlimited NS: unlimited
problem NV: unlimited NV: 3000 NV: unlimited NV: unlimited NV: unlimited
Running timea 15 minutes 2 minutes 1 minute seconds 1–5 minutes
Algorithm Heuristics-LNS Proprietary Custom heuristic Algorithm vary by Proprietary
heuristic scenario heuristic
Source: OR/MS-Today (2016).
Note: NS, number of stops and NV, Number of vehicles.
a
Solve problem with 50 routes, 1000 stops, two-hour hard-time windows.
8 N. RINCON-GARCIA ET AL.

the survey and that the current adoption rate in the industry is low (11% in companies with
less than 10 vehicles and 17% in companies with more than 10 vehicles) (DFT, 2010).
Additionally, in many cases it was stated that as a company policy, no information
could be provided to researchers. In total, 19 responses were gathered, 52.6% were
third-party logistics while 47.3% supported private operations where the core business
was not freight services. Characteristics of transport operations in the surveyed companies
are presented in Figure 3.
Among the main reasons to adopt CVRS are improvement of service level, reduction of
scheduler’s time and balance routes and control costs, see Figure 4. Surveyed companies
reported that the financial impact of congestion on service level had been greater than the
impact of driving time legislation, where more drivers are required to comply with driving
time legislation, see Figure 5.

Figure 3. Characteristics of transport operations in surveyed companies. Source: authors.

Figure 4. Reasons to adopt CVRS. Source: authors.

 TRANSPORT REVIEWS 9

Figure 5. Impact of new challenges in the industry. Source: authors.

Regarding software improvements, users were asked to score the functionality that, in
their opinion, required the most improvement using the following scale: 1 (large improvement
required) to 5 (excellent functionality). The reported capabilities that required the most
improvement were optimisation of routes minimising impact of congestion and accuracy of
information regarding vehicle restriction to certain zones and times with a score of 2.33.
The mean score of all capabilities was 3.09, accuracy of road representation obtained a
score of 3.16, where capabilities to provide support for the driving time directive was 3.35.
Support for different vehicle characteristics obtained the highest score (Figure 6), a capability
reported as a “must have” in CVRS according to Drexl (2012).

Figure 6. Users’ evaluation of CVRS capabilities according to the required improvement.

Source: authors.
10 N. RINCON-GARCIA ET AL.

Although the main vendors reported capabilities for routing with historical travel time and
to some extent, real-time traffic information, the surveyed companies stated the necessity for
improving these related capabilities. In the following section, a literature review is presented
in order to understand the available data, the state-of-the-art in VRP models that support the
core analytic tool used in CVRS that optimises the planning of routes, with a focus on time-
dependent models and lastly, the reported barriers in hardware are mentioned.

4. Literature review
4.1. Data
The two capabilities that require the most development are the optimisation of routes, mini-
mising the impact of congestion and the accuracy of information regarding vehicle restric-
tions to certain zones and times. Additional capabilities related to data are re-routing,
where real-time congestion might be required, and accuracy of road representation.
One issue reported by users is the quality of road characterisation, and the lack of accu-
rate, detailed road information tailored for freight vehicles. This has resulted in software
providing incorrect solutions, for example, inefficient planning and vehicles using unsuita-
ble roads (DFT, 2010). Gathering, processing and distributing data require policy and
resources, much of the information required by CVRS is obtained by local authorities
and initiatives have been put in place in order to openly share datasets from ITS.
There is a clear statement from policy-makers in the European Union to promote the
use of open data as it is recognised as the “raw material” for developments in information
products and services that can yield great economic and social benefits. The roadmap in
policy is based on adapting the legal framework, mobilising financing instruments and
facilitating coordination and experience (European-Commission, 2011). Therefore, it is
expected that app developers include all the benefits and capabilities of mobiles to disse-
minate useful information to users of transport systems (e.g. congested routes, better
journey planning, etc.), where much of the data can be reused into CVRS. The city of
Reading share the links of the road network and estimation in real time of road conditions
based on loop detectors in selected links.1 However, the freight industry requires a
number of road characteristics such as maximum height, weight and width restrictions,
road gradient, restricted turns and access, tolls and petrol stations. This level of detail is
expensive to obtain and the sources are often not clear in terms of ownership.
Different mapping products have come onto the market, some of which rely on the
users to update data, for example, Waze which has 140,000 volunteer map editors
(Gibbs, 2014). Although the freight industry requires highly detailed information, Waze
is an example of how to use and communicate driver’s knowledge. Regarding traffic pat-
terns, currently some applications provide travel time estimations even considering real-
time conditions, such as TomTom in some European Countries (Switzerland, the U.K.
and The Netherlands) (Cohn, 2009) and Google Traffic. Some of the initiatives deployed
to obtain traffic information are

(i) Crowdsourcing road congestion data: mobile technologies and GPS are used to get
information from multiple drivers in order to estimate network conditions (Marfia,
Roccetti, & Amoroso, 2013). Marfia et al. (2013) named Google Traffic as an
example of this initiative and suggested that data should be used carefully due to
 TRANSPORT REVIEWS 11

a lack of information about its accuracy. Privacy issues are also mentioned in the lit-
erature regarding user consent, data security and anonymisation (Laurila et al., 2012).
(ii) Automatic vehicle identification: registration plate recognition systems are used to
identify vehicles and estimate travel time between locations when the same vehicle
is identified in two different locations over a period of time (Dion & Rakha, 2006).
(iii) Spot speed measurement systems: speed detectors measure the speed of the
vehicles at the location of the sensor (e.g. loop detectors, radar and infrared technol-
ogies) (Dion & Rakha, 2006). However, travel time estimations might be inaccurate
due to the complexities involved in calculating speed based on the location of the
sensor and not considering traffic dynamics and queue evolution (Soriguera &
Robusté, 2011).

Traffic information is used to predict travel times on congested roads, it requires con-
tinuous data collection across the day in order to tailor the speed model (Kok et al., 2012),
gathering traffic information on minor roads and all the links on urban networks is a diffi-
cult task due to prohibitive investment cost in the case of loop detectors and cameras,
where crowdsourcing is emerging due to the fact that it leverages mobile network infra-
structure providing a cost-effective way to collect traffic data (Herrera et al., 2010).
Nevertheless, data managing is an expensive task that still is a matter of controversy
regarding who should be responsible for it, local councils, map providers, public agencies,
etc. (DFT, 2010). There are no definitive standards governing how road database infor-
mation should be collected, updated or distributed (DFT, 2008, 2010). Recommendations
to improve these issues are joined-up working between authorities, mapping providers
and software vendors; resolve compatibility issues between telematics, Sat-Nav and any
other IT freight solutions; and standardisation and collaboration (DFT, 2010). Cooperation
and standardisation between stakeholders is required in order to provide accurate data.
Still, there is the concern about price, should this information be free and government sub-
sidise it or it should come at a price. Nevertheless, government has started to address this
issues due to the possible benefits in users of transport systems (Shrestha et al., 2014).

4.2. The vehicle routing problem

The VRP is a generic name for a large set of problems involving the optimisation of routes
for a fleet of vehicles to serve a set of customers given a number of constraints that are
present in the logistics operation (Baldacci, Mingozzi, & Roberti, 2012), in this paper any
model modification of the VRP proposed by Dantzig and Ramser (1959) that represents
an additional restriction will be called a variant. It is considered a NP-hard problem due
to the complexity in consistently obtaining optimal solutions. Although algorithms that
guarantee optimality have been developed to solve problems involving a few hundreds
of customers for a few VRP variants for certain instances (Baldacci, Mingozzi, & Roberti,
2011; Kallehauge, Larsen, & Madsen, 2006), heuristics algorithms (procedures designed
specifically for a given problem that explore a limited search space (Toth & Vigo, 2001))
and metaheuristic algorithms (based on “intelligent” search strategies that allow a
robust exploration by local search or population search and specially designed procedures
to avoid local optima, for example, Tabu search, Simulated annealing and Genetic Algor-
ithm (Cordeau, Gendreau, Laporte, Potvin, & Semet, 2002; Glover & Kochenberger, 2003)
12 N. RINCON-GARCIA ET AL.

have been shown to be able to provide solutions for problems of considerable size and
“richness” (Laporte, 2009).
Since the 1960s, VRP variants have been one of the test problems for algorithm devel-
opment for models with combinatorial formulation (some of the decision variables can
either take values 0 or 1). Although heuristics are capable to provide fast solutions for
large instances and are easy to implement, the introduction of metaheuristics in the 90s
achieved higher accuracy in reasonable time for some well-studied variants. A common
framework to test solutions is to initially provide a set of instances for a certain variant
and propose new algorithms to be evaluated in terms of accuracy (the difference
between the derived results against the optimal value or the best-known value) and
speed (time required to get the reported solution) (Bräysy & Gendreau, 2005a, 2005b;
Cordeau et al., 2002). A list of some of the VRP variants that represent some of the con-
ditions found in the industry is presented in Table 3.
CVRS providers are not estranged from testing algorithms with theoretical research,
companies report the use of well-known instances to test their algorithms, for example,
Solomon (1987) (instances of 100 customers) and Gehring and Homberger (1999)
(instances up to 1000 customers) (Drexl, 2012). Additionally, in order to support the indus-
try, it is required to deal with models that consider multiple restrictions, called rich VRP
variants. A case is presented by de Armas and Melián-Batista (2015) for some companies
in the Canary Islands where the model deals with time windows, heterogeneous fleets,
site-dependent deliveries, dynamic insertion of new customers and prioritisation of custo-
mers (some services might be postponed until next day if the customer agrees), the pro-
posed solution algorithm is tested with different benchmark instances for different
restrictions in order understand its level of competitiveness.

Table 3. List of VRP variants.

VRP variant Acronym Description
Time windows VRPTW Each customer has a demand, a time window in which to be served
and a specific visit dwell time (Solomon (1987).
Pick-up and delivery VRPPDTW Goods are picked up at one location and delivered to other location
(Parragh, Doerner, & Hartl, 2008).
Heterogeneous vehicle fleets HVRP Different vehicle types with different costs and characteristics are
available (Taillard, 1999).
Site-dependent VRP (site-dependent SDVRP Certain customers can only be served by certain vehicles and vehicles
deliveries) might have different characteristics (e.g. size) (Pisinger & Ropke,
2007).
Multi-depot MDVRP Multiple depots are allowed (Pisinger & Ropke, 2007).
Periodic PVRP The planning period is extended to several days (Yu & Yang, 2011).
Truck driver scheduling problem VRTDSP Consideration of the set of rules that dictate the maximum number of
(driving time regulation) driving hours in a weekly planning horizon, resting periods and the
maximum driving time without a break in a working day (Goel &
Vidal, 2014).
Fleet size and mix pollution-routing FSMPRP A number of vehicle types are available and the main objective is the
problem reduction of vehicle fixed costs (depreciation, repair and
maintenance, tyres, etc.) and routing cost based on travel time
(wages) and fuel cost (distance, optimisation of speed across the
links and CO2 emissions) (Koç, Bektaş, Jabali, & Laporte, 2014).
Time-dependent TDVRP Consideration of travel time variations during periods of delivery in
order to represent congestion (Kok et al., 2012; Malandraki &
Daskin, 1992).
Dynamic DVRP Some of the information is revealed during the execution of the route
(Pillac, Gendreau, Guéret, & Medaglia, 2013).
 TRANSPORT REVIEWS 13

Metaheuristics algorithms for the VRP with time windows (VRPTW) can produce sol-
utions in a matter of seconds for 100 customers and minutes for 1000 customers (Vidal,
Crainic, Gendreau, & Prins, 2013), but there are a number of issues when implementing
metaheuristics for more complex variants that are exemplified in the VRP that accounts
for driving time regulation in the European Union. After its first formal formulation,
along with benchmarking instances and a solution method based on a large neighbour-
hood search (LNS) metaheuristic by Goel (2009), results with a different LNS tailoring pro-
posed by Prescott-Gagnon, Desaulniers, Drexl, and Rousseau (2010) obtained highly
improved results, namely a reduced number of vehicles (31.7%) and travel distance
(17.2%). Therefore, the issues that must be considered are the following:

(i) Tailoring solution methods for VRP variants is a complex activity that may yield low
accuracy when using non-exact algorithms (heuristics and metaheuristics).
(ii) Research in non-exact algorithms that do not provide benchmark instances or make
use of them to test results cannot guarantee accuracy and the industry must be
careful about their implementation.

An additional criticism of the theoretical research on VRP models is that in many cases,
they lack the ability to be reproduced as they are often highly tailored to the specific case
being studied. Some metaheuristic implementations have been tailored to work well in
specific test instances by tuning parameters so specifically as considering the best
random seed that provides high accuracy (Sörensen, 2015) and the industry requires algor-
ithms capable to consistently solve real-life problems with high accuracy without the
necessity of specific calibration of parameters according to the instance at hand. Further-
more, according to Drexl (2012, p. 60) “The last 0.1% in solution quality to be gained from
an additional complex algorithmic device are insignificant, since the data available in prac-
tice are never 100% accurate”. According to Cordeau et al. (2002) the characteristic of
“good” algorithms that interest the industry are speed, accuracy, simplicity (sufficient infor-
mation is provided for programmers to come up with a working code and algorithms
should be able to provide consistently high accuracy independently of the instance)
and flexibility (capability of the algorithm to accommodate constraints found in real-life
operation without deteriorating accuracy such as the variants shown in Table 3).

4.2.1. The time-dependent VRP

Gendreau, Ghiani, and Guerriero (2015) conducted the most recent literature review in
time-dependent routing problems (TDVRP). Some of the mentioned problems are the
“time-dependent point-to-point route planning” (which is obtaining the optimal path
between two locations in a road network) and the “TDVRP”. The challenge in the point-
to-point route planning relies on providing efficient algorithms online for the next-gener-
ation web-based travel information systems that require results in milliseconds or micro-
seconds. Although, it is required to use this problem to establish time-dependent travel
times in time-dependent variants, shortest paths might be determined in a pre-processing
phase prior to the execution of solving the actual schedule (Kok et al., 2012) due to the fact
that forecasted travel times are used and there is no need of online applications when
designing the routes for the following planning period. Additional VRP variants that con-
sider fuel consumption are also mentioned. Gendreau et al. (2015) highlight the
14 N. RINCON-GARCIA ET AL.

requirement of additional contributions of the operational research community in time-

dependent routing problems, where techniques for constant-speed classic network optim-
isation problems exist but it is required research for their time-dependent counterparts.
In the case of the TDVRP, the first formulation was introduced by Malandraki (1989) and
Malandraki and Daskin (1992) where travel time between customers depended on the
time of day, and the solution algorithm was based on heuristics and a cutting-plane algor-
ithm without the presence of time windows, for instances, up to 25 customers. A proposed
methodology to evaluate the impact of using constant versus time-dependent speed
models is to compute solutions with each model and compare scenarios considering
time-dependent travel times (Cattaruzza, Absi, Feillet, & González-Feliu, 2015). Fleischmann,
Gietz, and Gnutzmann (2004) proposed a real-world application based on traffic data from
Berlin, reaching to the conclusion that not using time-dependent models might underes-
timate travel time up to 10%. Likewise, Ehmke et al. (2012) found underestimation up to
20% in the area of Stuttgart. An additional issue is the reliability of links in the network, by
understanding the travel time distributions when sufficient data are available, it is possible
to create routes with low variation from the initial planned schedule (Ando & Taniguchi,
2006).
From the solution method, time-dependent models create new complexities for algor-
ithm design related to tailoring existing search strategies designed specifically for con-
stant-speed models. Common local search procedures require significant modification
as alterations within a route as part of the search process could potentially affect the feasi-
bility of the rest of the route. This might alter the departure times of subsequent visits to
customers and consequently modify travel times. Route evaluation is considerably more
computationally expensive with time-dependent travel times (Harwood, Mumford, &
Eglese, 2013). The TDVRP with soft time windows (where violation of time windows is
allowed) has been studied by Ichoua, Gendreau, and Potvin (2003) and Vidal, Crainic,
Gendreau, and Prins (2014), and the TDVRP with hard windows (where violation of time
windows is not allowed) by Donati, Montemanni, Casagrande, Rizzoli, and Gambardella
(2008), Balseiro, Loiseau, and Ramonet (2011), Kok et al. (2012) and Dabia, Ropke, Van
Woensel, and De Kok (2013). Common solution techniques are based on metaheuristic
algorithms, exact methods that guarantee optimality with mathematical approaches
have been studied by Dabia et al. (2013) using a modified set of the well-known instances
of Solomon (1987) for the VRPTW. Speed patterns in links between nodes were allocated
randomly and the solution approach was based on a pricing algorithm utilising a column
generation and a labelling algorithm. In trials, 63% of the 25 customer instances were
solved, 38% of the 50 customer and 15% of the 100 customer, solution might take
hours of computational time.
The first set of replicable instances to study accuracy of metaheuristic approaches for
the time-dependent VRPTW variant, providing all the information to reproduce the
instances as opposed to the use of random elements as in the research of Donati et al.
(2008) or Dabia et al. (2013), was proposed by Figliozzi (2008) and Figliozzi (2012) along
with a ruin and recreate algorithm as solution method. Rincon-Garcia, Waterson, and Cher-
rett (2016) recently proposed a LNS approach that improved average results in the 672 test
instances, namely reduction in the number of vehicles (4.15%), travel distance (10.88%)
and travel time (12.00%). Kok, Hans, Schutten, and Zijm (2010) introduced a VRP variant
with time-dependent travel times and a subset of the European driving time regulation
 TRANSPORT REVIEWS 15

applicable in urban or regional operations that might consider planning only one day
ahead, where proposed benchmark instances are an extension of Figliozzi (2008).
Rincon-Garcia, Velazquez, Waterson, and Cherrett (2015) extended the LNS approach to
this variant suggesting, in initial trials, the result improvements in reduction of number
of vehicles and distance over 10%.
The relationship between congestion, fuel consumption and CO2 emissions has been
studied by Maden, Eglese, and Black (2010) who compared the impacts of scheduling
vehicles with and without time-dependent travel times for a sample fleet in the south
of the U.K. The results suggested a reduction of 7% in CO2 emissions by considering
travel time dependency, increment of fuel consumption is related to low speed produced
by congestion. Furthermore, Bektaş and Laporte (2011) suggested that additional research
is required in analysis of heterogeneous fleets in congested environments due to possible
benefits of reductions in emissions and costs; where vehicles can be directed to less con-
gested roads even at the expense of longer distances specially in urban areas (Lin, Choy,
Ho, Chung, & Lam, 2014).
However, research in accurate algorithms for time-dependent variants is still scarce and
there are available instances for only a few variants. The industry and scientific community
should be careful about conclusions made with algorithms that might be inaccurate.

4.3. Hardware
CVRS might receive information from and make use of different ITS technologies, where
there are potential benefits from device interconnectivity, such as retrieving actual
driving times from the Tachograph in order to provide schedules for next day accommo-
dating driving time regulation, telematics and analysis of driver behaviour, downloading
detailed turn-by-turn schedule into the Sat-Nav for driver knowledge, etc. However, stan-
dardisation and collaboration are required to improve device connectivity (DFT, 2010).
An additional barrier mentioned in ITS related to routing and driving is accuracy of
vehicle location, where there is signal loss of GPS due to the “city canyon” effect; tall build-
ings in city environments, trees and hills might bounce radio signals from satellites and
derive inaccurate GPS location (Gong, Chen, Bialostozky, & Lawson, 2012). Although it is
a key issue in systems that require high accuracy, for example, driverless car, different sol-
utions have been created for the freight industry such as the integration of road charac-
teristics to estimate the position of the vehicle when no signal is received (DFT, 2010; Gong
et al., 2012).

5. Discussion
The concept of business logistics is more frequently recognised as a core capability that
enables companies to gain and maintain competitive advantage based on offering
better customer value, defined as the relation between perceived benefits and customer’s
costs, where effective planning and operation in logistics is able to improve both elements
and create value to customers (Christopher, 2010). When routing, it seems that companies
maintain these objectives, in the prioritisation of key routing factors reported by Rowell
(2012), meeting customers’ requirements and the minimisation of total cost (travel time
and travel distance) appear among the most important factors. Something according to
16 N. RINCON-GARCIA ET AL.

the prioritisation of reasons to adopt CVRS reported by companies in this research,

additional reasons are reduction of scheduler’s time and balancing workloads. Available
processor capacity allows the creation of feasible optimised routes that take under con-
sideration a range of restrictions in a fraction of the time that is required by schedulers.
Still, it is reported in the literature that human interaction is required to verify and
modify routes, Erdoğan, McLeod, Cherrett, and Bektaş (2015) exemplify it in a charity
organisation that adopted a prototype VRP model for bank charity collection and shop
delivery, transport manager yet has to modify routes in order to minimise the risk of
time window violations due to heavy traffic, balancing workloads, clustering, inclusion
of urgent requests, forced or delayed collections and unavailability of vehicles or staff.
When additional restrictions not supported in the model have to be considered or a
relaxed schedule is required, manual route modifications should be easily performed
and evaluated with an adequate CVRS interface design based on GIS (Cegarra et al., 2012).
The freight transport industry has some special characteristics, low margin profits (up to
3%) (FTA, 2015), a large number of operators with very few vehicles and a small number of
companies with a large number of vehicles (7% of operators controls 54% of the total
goods vehicles in the U.K.) (Davies, Mason, & Lalwani, 2007), and a large range of logistic
configurations to support different business models. Therefore the low CVRS adoption
rate, in many companies it does not support the specific type of operation or bring
enough benefits to compensate the investment of its implementation. Nevertheless, big
players are implementing and improving computerised routing such as UPS and its
ORION project (On-Road Integrated Optimisation and Navigation), it saves more than
1.5 million gallons of fuel, reduces the environmental impact of the operation and has
shown in preliminary tests the possibility to save millions of dollars. It will enable the
next generation of software capabilities by allowing customers to see online their incom-
ing deliveries and actively choose drop preferences and reroute shipment if necessary
(UPS, 2015), it is an example of how ITS has transformed the industry and will continue
to do so.

5.1. Academic theory, commercial VRP and industry needs

The freight industry commonly works on a low margin and costs are a key factor where
CVRS is recognised as a possible tool to adopt. The literature review shows the importance
of understanding the accuracy of algorithms, reported differences of quality of solution in
commercial software is up to 10% in accuracy in instances of small size (100 customers) for
constant-speed VRP variants, quality deteriorate with larger instances, and the industry
requires to solve instances up to thousands of deliveries.
The companies interviewed in this research considered that improving the capability of
optimising routes and minimising the impact of congestion were two of the most required
improvements in routing software. The impact of congestion has increased over the last 30
years, with the 101 largest U.S. cities reporting that travel delay had increased from 1.1
billion hours in 1982 to 4.8 billion hours in 2011 (Chang, Lee, & Choi, 2015). Figure 7
shows the variation in the level of congestion by time-of-day that affects the freight trans-
port in the U.S.
Although managerial solutions are available to mitigate the impact of congestion when
routing such as planning vehicle schedules with average travel times, it might lead to poor
 TRANSPORT REVIEWS 17

Figure 7. Variation in congestion by time-of-day. Source: US-Department-Of-Transport (2003).

solutions with missed deliveries and extra costs (e.g. more vehicles, duty time and dis-
tance) when compared to solutions based on the use of time-dependent VRP models
(Kok et al., 2012). Therefore, all the effort in providing highly accurate algorithms for con-
stant-speed models is lost when working in congested environments such as cities. Fur-
thermore, academic theory for time-dependent models that provides high accuracy is
quite scarce and is only available for a few variants (e.g. the time-dependent VRPTW
and its extension to cope with the European driving time regulation applicable in urban
or regional operations), and many of the theoretical variants for the VRP have not been
adequately studied when considering time-dependent travel times by providing special
methods to speed up the search process in order to deal with thousands of requests
and to cope with the multiple restrictions that have to be considered to provide
support to the industry.
Moreover, there is a change in the behaviour of final consumers, current retail trends
show that online sales represent 14% of all UK brick-and-mortar stores and e-commerce
and this is expected to rise up to 35% by 2020 (Javelin-Group, 2011; Visser et al., 2014)
where the last mile is considered as one of the most expensive, least efficient and polluting
sections of the entire logistic chain (Gevaers, Van de Voorde, & Vanelslander, 2011). The
study of time-dependent models has the potential to reduce the negative impacts of
freight transport and improve economical business conditions.

6. Conclusions
CVRS is used in the transport industry to improve service level by supporting planners in
the complex activity of vehicle scheduling. The literature review shows that a range of
technological breakthroughs has been implemented in this type of systems in order to
provide representation of road characteristics, online delivery status and vehicle location,
and traffic information in order to support software capabilities required by the industry.
However, it is important to understand the barriers found in real life when compared to
technological or theoretical concepts, which is intended in this research in order to
provide knowledge to software developers, researchers and stakeholders in order to ident-
ify the most relevant issues to contribute to the transport industry.
Congestion is reported as a major concern and sampled companies report that it affects
the operation in terms of cost increment and deteriorate service. Furthermore, companies
18 N. RINCON-GARCIA ET AL.

identified route optimisation under congestion as one of the two capabilities that require
the most development. The technological elements to improve planning under conges-
tion are traffic patterns and time-dependent models in order to provide reliable schedules
and both elements are reported as barriers; gathering information in all the links of the
network is a difficult task and although the VRP has received much attention from
researchers, time-dependent variants have not been yet subject of adequate algorithm
development. Nevertheless, there is traffic data at certain extend and evolution of oper-
ational research techniques is certainly capable to provide efficient solutions for this
kind of models along with other variants such as driving time regulation, where it is
required research in rich time-dependent models that make use of benchmark instances
to test accuracy.
The other capability that requires the most development is to provide access restric-
tions to roads according to type of vehicle in different zones and times, a reported
barrier in issues regarding data, responsibility of gathering, distributing and standardis-
ation is a matter of controversy, collaboration between government and private sector
is recommended to overcome it.
Models can be only as reliable as the data that are available, and the transport industry
recognises that further improvements are required to obtain reliable solutions based on
accurate data. It seems that cooperation across a range of stakeholders might offer
great benefits to the industry in order to cope with congestion and regulation.

Note
1. http://opendata.reading-travelinfo.co.uk/.

Disclosure statement
No potential conflict of interest was reported by the authors.

References
Ando, N., & Taniguchi, E. (2006). Travel time reliability in vehicle routing and scheduling with time
windows. Networks and Spatial Economics, 6, 293–311.
de Armas, J., & Melián-Batista, B. (2015). Variable neighborhood search for a dynamic rich vehicle
routing problem with time windows. Computers & Industrial Engineering, 85, 120–131.
Baldacci, R., Mingozzi, A., & Roberti, R. (2011). New route relaxation and pricing strategies for the
vehicle routing problem. Operations Research, 59(5), 1269–1283.
Baldacci, R., Mingozzi, A., & Roberti, R. (2012). Recent exact algorithms for solving the vehicle routing
problem under capacity and time window constraints. European Journal of Operational Research,
218(1), 1–6.
Balseiro, S., Loiseau, I., & Ramonet, J. (2011). An ant colony algorithm hybridized with insertion heur-
istics for the time dependent vehicle routing problem with time windows. Computers & Operations
Research, 38(6), 954–966.
Bektaş, T., & Laporte, G. (2011). The pollution-routing problem. Transportation Research Part B:
Methodological, 45, 1232–1250.
Bozkaya, B., Yanik, S., & Balcisoy, S. (2010). A GIS-based optimization framework for competitive multi-
facility location-routing problem. Networks and Spatial Economics, 10(3), 297–320.
Bräysy, O., & Gendreau, M. (2005a). Vehicle routing problem with time windows, part I: Route con-
struction and local search algorithms. Transportation Science, 39(1), 104–118.
 TRANSPORT REVIEWS 19

Bräysy, O., & Gendreau, M. (2005b). Vehicle routing problem with time windows, part II:
Metaheuristics. Transportation Science, 39(1), 119–139.
Bräysy, O., & Hasle, G. (2014). Software tools and emerging technologies for vehicle routing and inter-
modal transportation. Vehicle Routing: Problems, Methods, and Applications, 18, 351–380.
Cattaruzza, D., Absi, N., Feillet, D., & González-Feliu, J. (2015). Vehicle routing problems for city logis-
tics. EURO Journal on Transportation and Logistics, 1–29. doi:10.1007/s13676-014-0074-0
Cegarra, J., Gacias, B., & Lopez, P. (2012). Implications of technological changes in vehicle routing
interfaces for planners’ constraint processing. Human Factors and Ergonomics in Manufacturing
& Service Industries, 22(5), 468–480.
Chang, Y. S., Lee, Y. J., & Choi, S. B. (2015). More traffic congestion in larger cities? – Scaling
analysis of the large 101 US urban centers(March 25, 2015). Gachon GCCR Working Paper Series
No.15-5. Available at SSRN: https://ssrn.com/abstract=2584878 or http://dx.doi.org/10.2139/ssrn.
2584878.
Cherrett, T., Allen, J., McLeod, F., Maynard, S., Hickford, A., & Browne, M. (2012). Understanding urban
freight activity – key issues for freight planning. Journal of Transport Geography, 24, 22–32.
Cherrett, T., Shingleton, D., Norton, B., McLeod, F., Forey, C., Dickinson, J., … Norgate, S. (2015).
Developing a smartphone app to enhance Oxfam’s supply chain visibility. International Journal
of Logistics Research and Applications, 18(2), 155–167.
Christopher, M. (2010). New directions in logistics. In D. Waters (Ed.), Global logistics: New directions in
supply chain management (pp. 1–11). London: KoganPage.
Cohn, N. (2009). Real-time traffic information and navigation. Transportation Research Record: Journal
of the Transportation Research Board, 2129(1), 129–135.
Cordeau, J.-F., Gendreau, M., Laporte, G., Potvin, J.-Y., & Semet, F. (2002). A guide to vehicle routing
heuristics. Journal of the Operational Research Society, 53(5), 512–522.
Crainic, T. G., Gendreau, M., & Potvin, J.-Y. (2009). Intelligent freight-transportation systems:
Assessment and the contribution of operations research. Transportation Research Part C:
Emerging Technologies, 17(6), 541–557.
Dabia, S., Ropke, S., Van Woensel, T., & De Kok, T. (2013). Branch and price for the time-dependent
vehicle routing problem with time windows. Transportation Science, 47(3), 380–396.
Dantzig, G. B., & Ramser, J. H. (1959). The truck dispatching problem. Management Science, 6(1), 80–91.
Davies, I., Mason, R., & Lalwani, C. (2007). Assessing the impact of ICT on UK general haulage compa-
nies. International Journal of Production Economics, 106(1), 12–27.
DFT. (2005). Computerised vehicle routing and scheduling (CVRS) for efficient logistics. Author.
DFT. (2008). Transport statistics Great Britain. Author.
DFT. (2010). Integrated research study: HGV satellite navigation and route planning. Author.
Dion, F., & Rakha, H. (2006). Estimating dynamic roadway travel times using automatic vehicle
identification data for low sampling rates. Transportation Research Part B: Methodological, 40(9),
745–766.
Donati, A. V., Montemanni, R., Casagrande, N., Rizzoli, A. E., & Gambardella, L. M. (2008). Time depen-
dent vehicle routing problem with a multi ant colony system. European Journal of Operational
Research, 185(3), 1174–1191.
Drexl, M. (2012). Rich vehicle routing in theory and practice. Logistics Research, 5(1–2), 47–63.
Eglese, R., Maden, W., & Slater, A. (2006). A road timetable to aid vehicle routing and scheduling.
Computers & Operations Research, 33(12), 3508–3519.
Ehmke, J. F., Steinert, A., & Mattfeld, D. C. (2012). Advanced routing for city logistics service providers
based on time-dependent travel times. Journal of Computational Science, 3(4), 193–205.
Erdoğan, G., McLeod, F., Cherrett, T., & Bektaş, T. (2015). Matheuristics for solving a multi-attribute
collection problem for a charity organisation. Journal of the Operational Research Society, 66(2),
177–190.
European-Commission. (2011). Open data: An engine for innovation, growth and transparent govern-
ance. Retrieved from http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=COM:2011:0882:FIN:
EN:PDF
European-Commission. (2012). Road transport: A change of gear. Author.
20 N. RINCON-GARCIA ET AL.

Figliozzi, M. A. (2008, May). An iterative route construction and improvement algorithm for the vehicle
routing problem with soft and hard time windows. Paper presented at the applications of advanced
technologies in transportation (AATT) 2008 conference proceedings, Athens, Greece.
Figliozzi, M. A. (2012). The time dependent vehicle routing problem with time windows: Benchmark
problems, an efficient solution algorithm, and solution characteristics. Transportation Research Part
E: Logistics and Transportation Review, 48(3), 616–636. doi:10.1016/j.tre.2011.11.006
Fleischmann, B., Gietz, M., & Gnutzmann, S. (2004). Time-varying travel times in vehicle routing.
Transportation Science, 38(2), 160–173.
FTA. (2002). The London Lorry Ban: Exempt route network study. Retrieved from http://www.fta.co.uk/
export/sites/fta/_galleries/downloads/lorry_ban_case_study.pdf
FTA. (2015). Logistics report 2015: Delivering safe, efficient, sustainable logistics. About. Tunbridge
Wells: Author.
Gehring, H., & Homberger, J. (1999). A parallel hybrid evolutionary metaheuristic for the vehicle routing
problem with time windows. Paper presented at the proceedings of EUROGEN99.
Gendreau, M., Ghiani, G., & Guerriero, E. (2015). Time-dependent routing problems: A review.
Computers & Operations Research, 64, 189–197.
Gevaers, R., Van de Voorde, E., & Vanelslander, T. (2011). Characteristics and typology of last-mile
logistics from an innovation perspective in an urban context. In City distribution and urban
freight transport: Multiples perspectives (pp. 56–71). Cheltenham: Edward Elgar Publishing Ltd.
Gibbs, N. (2014). The app that lets you improve Sat-Navs. The Telegraph. Retrieved from http://www.
telegraph.co.uk/motoring/caraccessories/10909576/That-app-that-lets-you-improve-sat-navs.html
Glover, F., & Kochenberger, G. A. (2003). Handbook of metaheuristics. Dordrecht: Springer.
Goel, A. (2009). Vehicle scheduling and routing with drivers’ working hours. Transportation Science, 43
(1), 17–26.
Goel, A., & Vidal, T. (2014). Hours of service regulations in road freight transport: An optimization-
based international assessment. Transportation Science, 48(3), 391–412.
Golob, T. F., & Regan, A. C. (2001). Impacts of highway congestion on freight operations: Perceptions
of trucking industry managers. Transportation Research Part A: Policy and Practice, 35(7), 577–599.
Gong, H., Chen, C., Bialostozky, E., & Lawson, C. T. (2012). A GPS/GIS method for travel mode detection
in New York city. Computers, Environment and Urban Systems, 36(2), 131–139.
Haghani, A., & Jung, S. (2005). A dynamic vehicle routing problem with time-dependent travel times.
Computers & Operations Research, 32(11), 2959–2986.
Hallamaki, A., Hotokka, P., Brigatti, J., Nakari, P., Bräysy, O., & Ruohonen, T. (2007). Vehicle routing soft-
ware: A survey and case studies with finish data. Technical Report.
Harwood, K., Mumford, C., & Eglese, R. (2013). Investigating the use of metaheuristics for solving
single vehicle routing problems with time-varying traversal costs. Journal of the Operational
Research Society, 64, 34–47.
Herrera, J. C., Work, D. B., Herring, R., Ban, X. J., Jacobson, Q., & Bayen, A. M. (2010). Evaluation of traffic
data obtained via GPS-enabled mobile phones: The mobile century field experiment.
Transportation Research Part C: Emerging Technologies, 18(4), 568–583.
Ichoua, S., Gendreau, M., & Potvin, J.-Y. (2003). Vehicle dispatching with time-dependent travel times.
European Journal of Operational Research, 144(2), 379–396.
Ioannou, G., Kritikos, M., & Prastacos, G. (2002). Map-route: A GIS-based decision support system
for intra-city vehicle routing with time windows. Journal of the Operational Research Society, 53,
842–854.
Jarvenpaa, S. L., & Lang, K. R. (2005). Managing the paradoxes of mobile technology. Information
Systems Management, 22(4), 7–23.
Javelin-Group. (2011). How many stores will we really need? UK non-food retailing in 2020. White paper.
London: Javelin Group. Retrieved from http://www.javelingroup.com/wp-content/uploads/2014/
03/Javelin-Group_White_Paper_How_Many_Stores.pdf
Jensen, A., & Dahl, S. (2009). Truck drivers hours-of-service regulations and occupational health. Work:
A Journal of Prevention, Assessment and Rehabilitation, 33(3), 363–368.
Jung, H., Lee, K., & Chun, W. (2006). Integration of GIS, GPS, and optimization technologies for the
effective control of parcel delivery service. Computers & Industrial Engineering, 51(1), 154–162.
 TRANSPORT REVIEWS 21

Kallehauge, B., Larsen, J., & Madsen, O. B. (2006). Lagrangian duality applied to the vehicle routing
problem with time windows. Computers & Operations Research, 33(5), 1464–1487.
Keenan, P. B. (1998). Spatial decision support systems for vehicle routing. Decision Support Systems, 22
(1), 65–71.
Khattak, A. J., Targa, F., & Yim, Y. (2004). Advanced traveler information systems assessing the benefits
and costs of ITS (pp. 217–240). Boston: Springer.
Koç, Ç, Bektaş, T., Jabali, O., & Laporte, G. (2014). The fleet size and mix pollution-routing problem.
Transportation Research Part B: Methodological, 70, 239–254.
Kok, A., Hans, E., & Schutten, J. (2012). Vehicle routing under time-dependent travel times: The impact
of congestion avoidance. Computers & Operations Research, 39(5), 910–918.
Kok, A., Hans, E., Schutten, J., & Zijm, W. (2010). A dynamic programming heuristic for vehicle routing
with time-dependent travel times and required breaks. Flexible Services and Manufacturing Journal,
22(1–2), 83–108.
Laporte, G. (2009). Fifty years of vehicle routing. Transportation Science, 43(4), 408–416.
Laurila, J. K., Gatica-Perez, D., Aad, I., Bornet, O., Do, T.-M.-T., Dousse, O., … Miettinen, M. (2012). The
mobile data challenge: Big data for mobile computing research. In Proceedings of the Workshop on
the Nokia Mobile Data Challenge, in Conjunction with the 10th International Conference on
Pervasive Computing, pp. 1–8.
Lin, C., Choy, K. L., Ho, G. T. S., Chung, S. H., & Lam, H. Y. (2014). Survey of green vehicle routing
problem: Past and future trends. Expert Systems with Applications, 41, 1118–1138.
Maden, W., Eglese, R., & Black, D. (2010). Vehicle routing and scheduling with time-varying data:
A case study. Journal of the Operational Research Society, 61(3), 515–522.
Malandraki, C. (1989). Time dependent vehicle routing problems: Formulations, solution algorithms and
computational experiments. Thesis (Ph.D.). Northwestern University.
Malandraki, C., & Daskin, M. S. (1992). Time dependent vehicle routing problems: Formulations, prop-
erties and heuristic algorithms. Transportation Science, 26(3), 185–200.
Marfia, G., Roccetti, M., & Amoroso, A. (2013). A new traffic congestion prediction model for advanced
traveler information and management systems. Wireless Communications and Mobile Computing,
13(3), 266–276.
Montemanni, R., Gambardella, L. M., Rizzoli, A. E., & Donati, A. V. (2005). Ant colony system for a
dynamic vehicle routing problem. Journal of Combinatorial Optimization, 10(4), 327–343.
OR/MS-Today. (2010). Vehicle routing software survey. Retrieved from http://www.lionhrtpub.com/
orms/surveys/Vehicle_Routing/vrss.html
OR/MS-Today. (2014). Vehicle routing software survey. Retrieved from http://www.orms-today.org/
surveys/Vehicle_Routing/vrss.html
OR/MS-Today. (2016). Vehicle routing survey. Retrieved from http://www.paragonrouting.com/uk/
enewsletter/paragon-tachomaster-link
Paragon. (2009). Paragon/tachomaster link. Retrieved from http://www.paragonrouting.com/uk/
enewsletter/paragon-tachomaster-link
Parragh, S. N., Doerner, K. F., & Hartl, R. F. (2008). A survey on pickup and delivery problems. Journal
für Betriebswirtschaft, 58(1), 21–51.
Pillac, V., Gendreau, M., Guéret, C., & Medaglia, A. (2013). A review of dynamic vehicle routing pro-
blems. European Journal of Operational Research, 225, 1–11.
Pisinger, D., & Ropke, S. (2007). A general heuristic for vehicle routing problems. Computers &
Operations Research, 34(8), 2403–2435.
Ploskas, N., Ioannis, A., Jason, P., & Nikolaos, S. (2015). A collaborative spatial decision support system
for the capacitated vehicle routing problem on a tabletop display. In I. Linden, S. Liu, F. E. Dargam,
& J. Hernández (Eds.), Decision support systems IV-information and knowledge management in
decision processes (pp. 26–36). Cham: Springer International.
Prescott-Gagnon, E., Desaulniers, G., Drexl, M., & Rousseau, L.-M. (2010). European driver rules in
vehicle routing with time windows. Transportation Science, 44(4), 455–473.
Rincon-Garcia, N., Velazquez, A., Waterson, B., & Cherrett, T. (2015). Improved algorithms for routing
urban delivery vehicles. Paper presented at the logistics research network, Derby.
22 N. RINCON-GARCIA ET AL.

Rincon-Garcia, N., Waterson, B., & Cherrett, T. (2016). A hybrid metaheuristic for the time-dependent
vehicle routing problem with hard time windows. International Journal of Industrial Engineering
Computations [Online preview]. Retrieved from http://www.growingscience.com/ijiec/IJIEC_
2016_18.pdf
Rochat, Y., & Semet, F. (1994). A Tabu search approach for delivering pet food and flour in
Switzerland. Journal of the Operational Research Society, 45, 1233–1246.
Rowell, M. (2012). Improving statewide freight routing capabilities for sub-national commodity flows.
Washington DC: University of Washington.
Sadeghi-Niaraki, A., Varshosaz, M., Kim, K., & Jung, J. J. (2011). Real world representation of a road
network for route planning in GIS. Expert Systems with Applications, 38(10), 11999–12008.
Shim, J. P., Warkentin, M., Courtney, J. F., Power, D. J., Sharda, R., & Carlsson, C. (2002). Past, present,
and future of decision support technology. Decision Support Systems, 33(2), 111–126.
Shrestha, B. P., Hounsell, N. B., McDonald, M., Cre, I., McDonald, R., & Beasley, S. (2014). Transferability
and implementation issues of open data for application-based traveller information. 10th ITS
European Congress, Helsinki, FI, 16–19 June, p. 9.
Skabardonis, A., Varaiya, P., & Petty, K. F. (2003). Measuring recurrent and nonrecurrent traffic
congestion. Transportation Research Record: Journal of the Transportation Research Board, 1856(1),
118–124.
Solomon, M. M. (1987). Algorithms for the vehicle routing and scheduling problems with time
window constraints. Operations Research, 35(2), 254–265.
Sörensen, K. (2015). Metaheuristics-the metaphor exposed. International Transactions in Operational
Research, 22, 3–18.
Sörensen, K., Sevaux, M., & Schittekat, P. (2008). Multiple neighbourhood. Search in commercial VRP
packages: Evolving towards self-adaptive methods. In Adaptive and multilevel metaheuristics (pp.
239–253). Heidelberg: Springer.
Soriguera, F., & Robusté, F. (2011). Requiem for freeway travel time estimation methods based on
blind speed interpolations between point measurements. IEEE Transactions on Intelligent
Transportation Systems, 12(1), 291–297.
Taillard, ÉD. (1999). A heuristic column generation method for the heterogeneous fleet VRP. RAIRO –
Operations Research, 33(01), 1–14.
Tarantilis, C. D., & Kiranoudis, C. T. (2002). Using a spatial decision support system for solving the
vehicle routing problem. Information & Management, 39(5), 359–375.
Toth, P., & Vigo, D. (2001). The vehicle routing problem. Philadelphia: Siam.
UPS. (2015). ORION Backgrounder. Retrieved from https://www.pressroom.ups.com/pressroom/
ContentDetailsViewer.page?ConceptType=Factsheets&id=1426321616277-282
US-Department-Of-Transport. (2003). Final report – Traffic congestion and reliability: Linking solutions
to problems. Retrieved from http://www.ops.fhwa.dot.gov/congestion_report_04/chapter3.htm
Vidal, T., Crainic, T. G., Gendreau, M., & Prins, C. (2013). A hybrid genetic algorithm with adaptive
diversity management for a large class of vehicle routing problems with time-windows.
Computers & Operations Research, 40(1), 475–489.
Vidal, T., Crainic, T. G., Gendreau, M., & Prins, C. (2014). A unified solution framework for multi-attri-
bute vehicle routing problems. European Journal of Operational Research, 234(3), 658–673.
Visser, J., Nemoto, T., & Browne, M. (2014). Home delivery and the impacts on urban freight transport:
A review. Procedia – Social and Behavioral Sciences, 125, 15–27.
Weigel, D., & Cao, B. (1999). Applying GIS and OR techniques to solve sears technician-dispatching
and home delivery problems. Interfaces, 29(1), 112–130.
Yu, B., & Yang, Z. Z. (2011). An ant colony optimization model: The period vehicle routing problem
with time windows. Transportation Research Part E: Logistics and Transportation Review, 47(2),
166–181.