-
Palgrave Macmillan Journals and Operational Research Society are
collaborating with JSTOR to digitize, preserve and extend access to
The Journal of the Operational Research Society.
http://www.jstor.org
Map-Route: A GIS-Based Decision Support System for Intra-City
Vehicle Routing with Time Windows Author(s): G. Ioannou, M. N.
Kritikos and G. P. Prastacos Source: The Journal of the Operational
Research Society, Vol. 53, No. 8 (Aug., 2002), pp. 842-854
Published by: on behalf of the Palgrave Macmillan Journals
Operational Research SocietyStable URL:
http://www.jstor.org/stable/822912Accessed: 18-05-2015 03:33
UTC
REFERENCESLinked references are available on JSTOR for this
article:
http://www.jstor.org/stable/822912?seq=1&cid=pdf-reference#references_tab_contents
You may need to log in to JSTOR to access the linked
references.
Your use of the JSTOR archive indicates your acceptance of the
Terms & Conditions of Use, available at
http://www.jstor.org/page/ info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars,
researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information
technology and tools to increase productivity and facilitate new
forms of scholarship. For more information about JSTOR, please
contact [email protected].
This content downloaded from 202.43.95.117 on Mon, 18 May 2015
03:33:35 UTCAll use subject to JSTOR Terms and Conditions
-
Journal of the Operational Research Society (2002) 53, 842-854 ?
2002 Operational Research Society Ltd. All rights reserved.
0160-5682/02 $15.00
www.palgrave-journals.com/jors
Map-Route: a GIS-based decision support system for intra-city
vehicle routing with time windows G Ioannou*, MIN Kritikos and GP
Prastacos Athens University of Economics and Business, Athens,
Greece
This paper presents a Decision Support System (DSS) that enables
dispatchers-schedulers to approach intra-city vehicle routing
problems with time windows interactively, using appropriate
computational methods and exploiting a custom knowledge base that
contains information about traffic and spatial data. The DSS, named
Map-Route, generates routes that satisfy time and vehicle capacity
constraints. Its computational engine is based on an effective
heuristic method for solving the underlying optimization problem,
while its implementation is developed using MapInfo, a popular
Geographical Information System (GIS) platform. Map-Route provides
very efficient solutions, is particularly user- friendly, and can
reach answers for a wide variety of 'what if' scenarios with
potentially significant cost implications. We have implemented
Map-Route in an actual industrial environment and we report on the
experience gained from this real- life application. Journal of the
Operational Research Society (2002) 53, 842-854. doi:l0.
1057/palgravejors.2601375 Keywords: vehicle routing;
distribution/logistics; decision support systems; heuristics;
GIS
Introduction
The vehicle routing problem with time windows (VRPTW) arises in
a variety of pick-up and delivery applications and can be described
as the design of optimal delivery/ collection routes from one or
several depots to a number of customers, within a pre-specified
time window, at mini- mum cost. Many papers in the literature have
addressed the VRPTW problem, and substantial research effort has
been devoted in developing efficient algorithms for solving a
variety of VRPTW problems. We refer to Bodin,l Laporte2 and
Gendreau et al3 for surveys of the VR literature and appropriate
pointers to relevant research efforts.
Since the mid-1980s, significant work has been performed in
developing computerized routing software systems.4 Examples are:
Geo-route,5 Fleet-Manager,6 micro-ALTO,7 Greentrip Toolkit,8
MACS-VRPTW,9 Dynamic Route Guidance,'0 and DRIVE." Apart from
general and/or dynamic VRPTW software, there have been several
industry specific approaches such as the ones summarized in Camp-
bell and Langevin12 for roadway snow and ice control, and Road-net,
Truck stops, and Micro Vehicle Plan, in the soft drink industry.13
Commercial software is also available for various applications (see
eg, http://www.geocities.com and http:
//www.paragon-software.co.uk).
*Correspondence. G Ioannou, Management Sciences Laboratory,
Gradu- ate Program in Decision Sciences, Department of Management
Science and Technology, Athens University of Economics and
Business, 8th Floor, 47A Evelpidon Street and 33 Lejkados Street,
Athens 113-62, Greece. E-mail: ioannougaueb.gr
The majority of the systems to-date have been of help to
enterprises; however for most of them a number of draw- backs have
been reported: (a) they are quite expensive, thus not preferred
solutions for Small and Medium Enterprises (SMEs)-the majority of
users; (b) they are based on proprietary software as opposed to
popular GIS platforms with standardized user-interfaces, effective
personnel train- ing, and guaranteed maintenance and system
upgrades; and (c) they need to incorporate more realistic
assumptions, and to improve solutions graphically. As a result,
distribution, pick-up and delivery SMEs still need suitable tools
for supporting complex decisions related to route planning, in
order to provide high level service to their customers and optimize
their resources.
The objective of this paper is threefold: first, to propose a
framework for addressing VRPTW for intra-city networks in a
user-friendly and effective manner via efficient solutions of the
underlying optimization problem. Second, to develop a prototype DSS
based on: (i) a popular GIS platform; and (ii) an optimization
method coupled with a knowledge base. And third, to demonstrate the
applicability of the approach through the results obtained from the
implementation of the DSS to an actual industrial environment. The
proposed DSS can assist logistics operations in a number of ways,
eg: (a) enhance daily operational tasks of dispatchers-schedu-
lers; (b) provide flexibility in solving VRPTW by generating
alternative solutions and reformulating problem conditions (eg,
editing the underlying transportation network, adding or removing
customers, defining alternative scenarios etc.), while keeping a
good 'eye' on the geographical reality of
This content downloaded from 202.43.95.117 on Mon, 18 May 2015
03:33:35 UTCAll use subject to JSTOR Terms and Conditions
-
G loannou et al-Map-Route: a GIS*based decision support system
843
the problem; and (c) offer interactivity, ie, allowing users to
employ visual techniques to formulate-reformulate problems and
derive solutions that can be easily implemented.
The remainder of the paper is organized as follows: first, we
present the architecture of our Decision Support System (DSS) and
its constituent elements. Then we provide the overall framework of
Map-Route and identify the interaction between all of its
components. The report on an industrial application follows, and
the conclusions of our work are finally presented.
Map-Route architecture
Map-Route is specifically designed for vehicle fleet routing for
deliveries within a compact large city street network, rather than
general VRPTW It consists of four basic com- ponents, which are
presented in detail below.
Databases of Map-Route
Map-Route's spatial database includes a digitized map with all
relevant locations (depot and customers) and underlying network
(streets, roads, intersections, etc.). The customer database
includes, for each customer, the node identification number, the
demand, and the time window restrictions and service time. The
nodes database includes, for each node, the identification number
and coordinates. Finally, the street database includes, for each
street segment, name, length, and address ranges for both sides.
The data files of Map- Route can be changeable or permanent. The
former relate to the properties of the underlying transportation
network and the coordinates of the depot location node. The latter
may be modified by the scheduler through our DSS by, eg, inputting
a new scenario via tables, entering new customers into a given
problem, removing customers from a scenario, input- ting a new
scenario using the map under consideration, and entering data
related to the problem using the browser table.
Computational engine of Map-Route The computational engine of
Map-Route can include any heuristic or mathematical programming
method for solving the VRPTW The selected method is very important
since it determines the applicability of the solution scheme in
real- life situations. The key factor for the appropriate selection
is the efficiency of the method and its ability to provide in very
short times high-quality solutions. In our approach, we use
IMPACT,14 the basic steps of which are:
Algorithm IMPACT Step 0: Initialization. Read the number of
customers, the
vehicle capacity, the inter-customer and depot- customer
distances (or times, routing costs) and
the earliest and latest service times (time window) for each
customer.
Step 1: Select a 'seed' customer to start a route, finding the
farthest customer from the depot. If there is no non-routed
feasible customer to start a route, go to Step 6.
Step 2: Find the feasible non-routed customer u that minimizes a
composite criterion Impact(u), which includes functions of the
relationship between the arrival time to customer u and the lower
bound on the service time of u, the impact of customer's u
insertion on non-routed custo- mers, and the impact of customer's u
insertion on customers already routed within the route under
construction. The search procedure is as follows:
Step 2a: Examine all possible feasible insertions of custo- mer
u into the current route. For each feasible insertion, calculate
the criterion function Impact(u). Select the insertion location
that results in minimum Impact(u) for this customer.
Step 2b: Repeat Step 2a for all feasible non-routed
customers.
Step 2c: Select customer u with minimum Impact(u). Step 3.
Insert the selected customer u, to the best inser-
tion location on the current route (see Steps 2a and 2c). Update
the route and set u as a routed customer.
Step 4: If there are non-routed customers that are feasi- ble
for insertion into the current route, return to Step 2; otherwise
proceed to Step 5.
Step 5. If all customers have been scheduled, terminate.
Otherwise, go to Step 1 initiating new route.
The algorithm terminates by providing the number of routes
(equal to active vehicles), the customers that are assigned to each
vehicle, the sequence in which customers are visited, and the total
time-distance-cost of the solution. IMPACT is very efficient and
provides results comparable to meta-heuristics at a fraction of the
computational effort. For a detailed description of IMPACT and the
computational tests that support its effectiveness, the reader is
referred to Ioannou et al,14 while Table 1 provides a comparison of
IMPACT with several heuristics (I-1, PARIS, HE) and meta- heuristic
(GRASP, Tabu search-TABU-A and RTS, and genetic
algorithms-GENEROUS-20) methods to illustrate its efficiency ('*'
indicates that IMPACT outperforms other methods for the classical
data sets RI, Cl, RC 1, R2, C2, and RC2 of Solomon15).
User-interface of Map-Route Map-Route's user-interface has been
developed using the Map-Basic programming language and is based on
pull- down menus to provide functionality related to solution
methods, problem initialization (eg, defining the speed of
This content downloaded from 202.43.95.117 on Mon, 18 May 2015
03:33:35 UTCAll use subject to JSTOR Terms and Conditions
-
844 Journal of the Operational Research Society Vol. 53, No.
8
Table 1 Comparison between literature heuristics and new
heuristic on average number of routes
Data set I-I PARIS HE GRASP TABU-A RTS GENEROUS-20
R1 * * * --* * R2 * * * * * * * C1 * * * * * * * C2 * * * _ RC1
* * * * * * RC2 * * * *
vehicles), data input, formation of local networks, design of
accurate vehicle routes, etc. Furthermore, it allows the display of
spatial maps allowing the user to zoom on a part of the map and
evaluate the suggested solution or generate alternative solutions
using logical inference. The user manuals of MapInfo16 provide all
the necessary functionality information.
Knowledge base of Map-Route
Deriving solutions that follow actual road networks and
complying to traffic patterns that favour main city arteries, large
streets and roads with light traffic and open structure is a very
difficult task, especially when optimization app- roaches are
employed. The latter are very sensitive to the route segments that
characterize the underlying road-street network and cannot handle
logical attributes such as those described above. The solutions
they produce comprise segments that may not be feasible in actual
route planning, or may not be preferable to drivers. The above
problems may be alleviated either by direct user interaction, ie,
changes in the route structure that are performed manually by
experienced users, or by appropriate knowledge bases that capture
the logic in which such changes are made. The first approach
requires high level of user involvement in the solution process and
is very time consuming for large problems, since the user has to
examine all parts of the network and make the necessary changes.
The latter approach requires a strong set-up phase in which roads
are coupled with specific prioritization attributes and routes are
examined using a knowledge base containing all these attributes or
additional information concerning preferences.
The structure of the knowledge base we propose is simple: Rules
and attributes are assigned to road segments and inference logic is
designed in order to transform an optimisation solution into a
feasible-preferable solution with minor cost implications. The
rules can have, eg, the following forms relevant to street
characteristics, time- related traffic and date peculiarities,
respectively:
'IF MULTILINE x IS { main, regular, narrow ) street then label =
{ PREFERRED, NONE, LOW )'
'IF MULTILINE x IS traffic loaded at time t then label =
LOW'
'IF MULTILINE x IS non-preferable on date then label = LOW'
Labels characterize the priority for using a particular route
segment (ie, segments with priority LOW will be used only when
necessary to guarantee the connectivity of a sub-network). The
rules are exhaustive for all appropriate road segments and time-
and date-related information, and are implemented in conjunction
with the Map-Basic routines using the experience of drivers and
planners-schedulers. The priorities associated with road segments
are directly used when forming a transportation network over which
vehicles are to be routed.17
Solution framework The proposed solution framework follows a
typical four-step process: (a) solve an approximation of the VRTPW
using Euclidean distances; (b) break the region down into sub-
networks, each corresponding to a vehicle route, and generate a
travel path over each sub-network using shortest paths between
customers, while preserving the order of customers in routes; (c)
modify the solution via the knowledge base rules to better approach
the actual vehicle paths; and (d) perform manual modifications, if
necessary. The solution process iterates among these steps until a
'good' solution is obtained. Subsequently the user accepts the
solution or modifies some of its attributes in order to provide the
final set of routes. The user can also modify problem parameters
and reapply the four steps if the system proposed solution is not
satisfactory. Visualization can play a critical role in this
process, and the GIS platform is very helpful in this
direction.
Figure 1 provides an overview of Map-Route's infrastruc- ture
and organization. MapInfo is the centre of the approach. Databases
include digitized maps and data concerning customers and depot. The
knowledge base is represented as a database that is external but
connected to MapInfo via an appropriate API. The user can interact
with the spatial data and also provide adjustnents graphically to
the solu- tions provided by the computational engine. In Figure 1,
lines connect sequentially evoked components, while arrows provide
the direction of each sequence.
Map-Route phases
As mentioned before, Map-Route involves four interacting phases.
During the first phase, the VRPTW is solved for a network where the
customers and the depot are connected
This content downloaded from 202.43.95.117 on Mon, 18 May 2015
03:33:35 UTCAll use subject to JSTOR Terms and Conditions
-
G loannou et al-Map-Route: a GIS based decision support system
845
Algorithms (Computational -4
Engine)
7\ . _ / D~~~~~~~~~~~~~epot / Pick-u
g t ze Mp10 APNO Deie y
Figure 1 The architecture of Map-Route.
with straight lines. This enables the solution of problems with
a large number of customers, in short computational times. Standard
MapInfo tools estimate the Euclidean distances between customers
and between customers and the depot.
During Phase 2, for every route of Phase 1, a connected
sub-network is constructed by selecting areas (road sections)
adjacent to the Euclidean routes. Sub-networks are adjusted through
MapInfo tools that use proximity criteria for actual route segment
inclusion. Furthermore, distances are reeval- uated based on the
priority rules of the knowledge base. On the new networks, accurate
routes can be determined according to the following steps: Step 1:
Renumber the nodes of the selected sub-network
(assigning '1' to the depot) Step 2: Run the Floyd's Shortest
Path algorithm'8
between customer locations (stops) and depot, in the same order
as in the solution of Phase 1
Step 1: Display the final accurate route on the original map The
above steps can be repeated for each sub-network, leading to
accurate routes that incorporate actual road segments. Note that in
Phase 2, the distance between customer locations is increased,
since the multi-lines of Phase 2 replace the straight lines of
Phase 1, and violations of customers' time windows may occur. This
problem can be addressed by: (a) selecting sub-networks more
adjacent to 'specific' trips; and (b) tightening time windows of
specific customers and iterating the whole process of the two
phases. Both approaches have been examined, and the results showed
that multiple iterations with tighter time windows are
preferable.19
It is important to note that the two-step approach of
determining Euclidean routes and transforming them into actual
street segment-routes may not be necessary or even
efficient for general VRPTW, especially in the case of large
inter-city routing with significant obstacles and constraints,
where severe problems may arise. Nevertheless, for a compact
intra-city network such as the one we are handling via Map-Route,
the approach can smoothly work.
In Phase 3 the knowledge base rules are evoked and the solution
is transformed to approach better the road segments employed by the
vehicles. This is accomplished by feeding the solution to the
knowledge base via the MapInfo inter- faces. When Phase 3 is
completed, the actual road network is determined based on proximity
criteria and preferences residing within the rule constructs.
Finally, Phase 4 is a pure user-driven phase. The user interacts
with MapInfo via the solution of Phase 3 and provides final
adjustments necessary to derive the schedule of each vehicle.
Figure 2 illustrates the four-phase approach inherent in the
Map-Route logic. Note that this is a discrete and time-including
representation of the Map-Route archi- tecture of Figure 1. The
elements of each Phase are grouped in shaded boxes, while the
sequence of the approach is depicted through the directional arcs
that connect databases, applications, results and user
adjustments.
Before we proceed to the implementation aspects of Map- Route,
we should mention that the optimal solution to intra- city routing
problems that we consider in this paper might include more than one
daily trip per vehicle. However, the common policy in all
distribution companies we have interacted with was to load only
once the vehicles at the warehouse and perform all remaining
activities the rest of the day. Furthermore, reaching near-truck
load per vehicle was a key performance indicator. Thus, we did not
proceed in exploiting this potential cost-saving application and
constrained our DSS into single daily loading and single routes per
day per vehicle. Nevertheless, an extension to
This content downloaded from 202.43.95.117 on Mon, 18 May 2015
03:33:35 UTCAll use subject to JSTOR Terms and Conditions
-
846 Journal of the Operational Research Society Vol. 53, No.
8
::
: E g : ; iS) LE ; y i i :: i y :: E L:E g - yiiLE;iLLi i...L
E:;; .........
0040-i00ll l--0000-000000000000000000-;0
t0000St000tt--0000000000t- ti-00f000000000ftPUser40:EW Q E Ei 24i D
EdQj i ER iED di R[i TiE iS i -iE Ei d E Ei i QiSEi EA d ju s tm e
n t d-253! i-iEES E Problem 0..::1X000 7 ... X10. S};000144XXa
ES000041001;S4.......S:XfS
0000StS1:X0-S-;::.:00000000A10X,fSjt400
. S . f. i MA.INFOf.Sf00tEfff :j.:fLiff4Jj 40E4W Et..iiE.
tEEttt:;i,;.ST i.itSN4X!.;;;;;. tidjf;; i g.f; i;;i; i;L .
LLiLa;;;itiz e d ,.. ap;LiLL0;st.L . .
i E . i TE R -iE E iS ;-.9- i - . -iSTi.9: - E g lD: 9 :D i; .
iE- E i~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.. .. . .. n
_; ,- - ---IMPi:::XX-XXA -- CT: j ::i:;e;XX- ::- XXESS
rHeuristica us:::104i000 40:00 0d 1 0X0:
;X0t:;---40000tV;f;X-F100000000:00f00410 04000t
i SC111 .Eif. aTf F l y '
S ho t est7.0 d . t. iLLL,iiESLLSS . St00tttl-000iTt000
tlltt00-0 iAf,|;ltlt;-fk0-0 -,S,,jttiSl0000 f-0.-tk--00-0.000-.00
......Results... 0--0.
_~~~~~~~~~~~~~~~~~~P MAPINFOhm
t t Digitized Maps
0~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.....
SolHurition netw Foy'sSortest basesut
.giiii . .. . .S .. L tE
::~~~~~~~~~~~~~~Fgr 2 Th Map;$ -Route: solutio flgow.tm
Map-Route is possible through reduced time available per day, a
factor that could be interactively modified within the DSS to
produce various solutions.
Map-Route implementation
Map-Route has been implemented on a Pentium PC. The core of DSS
has been written in Map-Basic, and the algorithms for the VRPTW in
Fortran, appropriately inte- grated into MapInfo. The knowledge
base uses a Lisp inference engine and is also integrated with
MapInfo. The solution provided by Map-Route is depicted on a real
city map, and generates informative output for the vehicle driver,
while enabling the evaluation of alternative routes. It is
important to restate the significance of user involvement in the
solution procedure. No matter how, extensive and complete the
knowledge base is, or effective and compre- hensive the
optimization solution is, the final set of routes is either
accepted or modified by the scheduler-planner, whose experience and
flexibility in dynamic daily adjustments of the problem parameters
is irreplaceable.
Industrial application
Background
The company for which Map-Route was developed is a wholesaler
and logistics service provider that supplies multi- ple packaged
goods and beverages to a large number of
local small supermarkets and other small retail outlets
throughout the Central Athens area, in Athens, Greece on a daily
basis. The company operates its own small-vehicle fleet from a
central warehouse located at Pireus (noted as PIRAIVS at the map of
MapInfo provided later in this section) Street, a main street
connecting Pireus to Omonia Square in the centre of Athens. The
company owns 26 delivery vehicles, which are operated by certified
drivers. The overall fleet size though, necessary for satisfying
all customers was approximately 35 vehicles, before the imple-
mentation of our DSS; thus, the company employed vehicles owned by
individuals on a need-basis, a fact the created additional costs
and resulted in severe problems with respect to quality of customer
service and adherence to order fulfilment goals.
The number of customers varies from 435 to 680, depend- ing on
the day of the week and the period of the year (higher number of
demand points during the summer season, when additional points of
sale are open to service the large tourist population that visits
Athens). The customers are dispersed throughout central Athens, and
are located at main arteries of the city (eg, Panepistimiou or
Stadiou Street) as well as at small streets near the archaeological
sites and particularly vibrant or densely populated city
neighbourhoods. Figure 3 illustrates the distribution of the
customer set on the map of Athens (circles), and the depot
(square).
Since a just-in-time approach is promoted as the compe- titive
advantage of the company, the replenishment of goods follows daily
orders from each customer; these orders may
This content downloaded from 202.43.95.117 on Mon, 18 May 2015
03:33:35 UTCAll use subject to JSTOR Terms and Conditions
-
G loannou et al-Map-Route: a GIS-based decision support system
847
Figure 3 Customer distribution and depot location on the map of
Athens.
be zero for some Stock Keeping Units (SKU) in a day.
Nevertheless, the overall demand is relatively constant, apart from
peak seasons, and especially during the summer where beverage
consumption increases significantly. The daily iterative operations
start with customer orders, which are fina- lized every evening,
and can be satisfied by the inventory held at the warehouse.
Inventory availability is guaranteed by the large safety stock held
for each SKU. The customers are geographically dispersed within a
distance radius that allows for demand to be satisfied through
daily deliveries, as shown in Figure 3. In addition, the time
interval during which the delivery has to take place (time window)
is also known (fixed for each customer according to a
contract).
The delivery process is performed as follows. Products are
loaded on vehicles at warehouse docks up to (or sometimes below,
according to customer requests) capacity and they are transported
to the customers' locations. At each location, quantities that
equal customer demand for each SKU are unloaded, and paper work
(shipping documents, bills and invoices) is filled and exchanged;
this takes approximately 5 min. Then, vehicles travel to subsequent
customers where the process is repeated, until all deliveries have
been performed and return to the depot for the following daily
cycle. It is important to note that before Map-Route's
implementation, the sequence in which a vehicle visited customers
was not determined when loading at the depot; drivers responsible
for a particular area-customer set were
making sequencing decisions. This had a significant effect of
the compliance to time windows, and affected cost, customer
satisfaction and quality of service.
Map-Route set-up
To generate the problem within MapInfo, we have started with
appropriate maps of the Central Athens area, and created 5137
node-objects for the 8231 road segments of the underlying map that
model approximately 3000 different streets and covering almost 500
km of road network, using the configuration tools of MapInfo. This
initialization is required for any subsequent task. Figure 4
provides a zoomed view around the depot of the road network model-
led in MapInfo for the application. To input the customer location
coordinates and the data concerning time windows, we used
appropriate files for data entry into MapInfo. Furthermore, we have
created special MapInfo screens to be available to the planners for
adjustments, deletions and additions of customers and time windows.
Figure 5 provides a sample screen that was constructed to give
multiple points of access to the user (both graphical and in
tabular form).
The road segments (black lines in Figure 3) were char- acterized
as 'preferred', 'unacceptable' or 'non-labelled', and this
information was included in the knowledge base. Filling up the
knowledge base with rules and priorities was the most daunting and
time-consuming task of the system
This content downloaded from 202.43.95.117 on Mon, 18 May 2015
03:33:35 UTCAll use subject to JSTOR Terms and Conditions
-
848 Journal of the Operational Research Society Vol. 53, No.
8
Figure 4 The zoomed map around the depot.
YA~ ~ ~~ 1844' 23.73225 37.O70081 0 00. 00 '23:50 189 23.733177
37.972683 - 0.0 35
1841 23.732472 37.973121 0 00XJ 0 2 5
~~~~~~ 2g1 ~~~~~~~~~~~~~~~~~~~~1642 23,730Q8M , 3. 2 0 '010
23
Lo - ~~Figre-5Dat ma ipuaion------screen. -----------
This content downloaded from 202.43.95.117 on Mon, 18 May 2015
03:33:35 UTCAll use subject to JSTOR Terms and Conditions
-
G loannou et al-Map-Route: a GIS*based decision support system
849
set-up process. We used data from the Greek Ministry of
Transportation and Communications concerning traffic patterns and
time-info for various dates of the year and times of the day.
Furthermore, we interviewed the drivers of the company for routing
preferences and considered their answers for labelling road
segments. Drivers were requested to provide the most commonly used
streets and estimates of travelling times at these segments during
peak and off-peak hours. Finally, for each 'unacceptable' road
segment, we run a special MapInfo procedure to derive via proximity
measures a 'preferable' corresponding road segment, and included it
in the knowledge base. It is important to note that the experience
of drivers conflicted some times with official data; however the
company's management insisted on adher- ing to drivers preferences,
as more reliable information concerning actual routing paths. Apart
from the initial set- up of the knowledge base, we have provided
screens within MapInfo, which allow users to adjust the labels
according to new realities, as the system life cycle evolves, or on
a daily basis, in line with expectations concerning congestion,
road blocking (strikes and marches in the centre of Athens is
commonplace), etc.
The computational engine of Map-Route, ie, IMPACT, was
integrated in the MapInfo menu. For the particular instances in the
industrial case, IMPACT required less than 30 s to terminate (for
the larger examples of more than 600 customers). We have also
incorporated Floyd's algorithm within the DSS. Floyd's
implementation uses dynamic tables in order to provide fast the
optimal solutions; for the particular instances in the industrial
case, the algorithm took less than 40s to terminate, even in cases
where the sub-network included a large number of node-objects due
to alternative route segments induced by the knowledge base.
The DSS in operation At the start of a shift, customer demand is
already into the system and Phase 1 of Map-Route is initiated. The
result is the sequence of customers visited by each vehicle based
on Euclidean distances, which are automatically calculated by
MapInfo. Euclidean routes appear on the screen with lines
connecting customers and depot; such a screen from the actual
application is provided in Figure 6. Note that the user can make
adjustments to time windows, demand and custo- mer attributes
(existence, location, etc.) before running IMPACT, if necessary,
through appropriate selections in the MapInfo menu (that invoke the
previously discussed screens).
Subsequently, the user proceeds to the second phase of Map-Route
to determine the actual road path of each vehicle using the
shortest paths on the real road network. The procedure is repeated
for each vehicle and is as follows: An initial sub-network is
formed through the knowledge base rules of proximity; this
sub-network, which includes various road segments, is expanded or
adjusted by the user that can
include additional segments or remove some segments based on
experience and daily data. Given the sub-network, a routine
incorporated in MapInfo produces the necessary shortest path
matrix. Figure 7 presents a sample sub-network associated with one
route of the Euclidean solution of Phase 1 presented in Figure 6
(includes all route segments depicted by the thick lines).
Given the distance matrix, the next step is to apply Floyd's
algorithm to determine the actual vehicle paths by invoking a
resident MapInfo routine for calculating shortest paths and
displaying the results on the MapInfo interface. Figure 8 provides
the actual road path for the sub-network of Figure 7, and Figure 9
a zoomed view. The procedure is repeated for each initial Euclidean
route of Phase 1. This loop constitutes the most time consuming
part of the application, since it is directly linked to the number
of vehicles employed. If this number remains at the present level
(ie, order of 30 vehicles), then it is possible to complete the
routing procedure and derive schedules for each vehicle in less
than 1 h. This time is acceptable, and allows the company to
smoothly employ the DSS. However, future plans include the addition
of several more SKUs and customers-locations, a fact that would add
further delay to the application. Thus, we were asked to automate a
combined Phase 1-2 of Map-Route. This provided full solutions
(actual road networks for all vehicles) that could be further
examined and improved a posteriori by the users, if necessary. The
automated procedure allowed the completion of the daily tasks in
less than 0.5 h. However, planners who desired their direct
intervention and drivers who felt that their flexibility was
compromised did not deem the total automation appropriate. Thus,
the operational version of Map-Route runs with individual route
construc- tion and adjustments, and performs the routing procedure
sequentially for each vehicle.
The results of Map-Route are provided to the drivers, who are
requested to follow the prescribed routes (customer sequences and
road segments) in their daily delivery sche- dule. Appropriate
forms that include the sequence of custo- mers to be visited and
the expected time of arrival at each customer's location are
employed to check the delivery schedule; customers are required to
sign the form when paperwork is exchanged.
Results and discussion of the application Map-Route was deployed
to two PCs of the company, lo- cated at the office of the
warehouse. The DSS was stand- alone, ie, it was not connected to
any other Information Systems (eg, the warehouse and inventory
management sys- tems). Two planners, who were involved in all
stages of the finalization-deployment of the DSS, were responsible
for data entry and running Map-Route on a daily basis. The sys- tem
was tested using demand data from previous time peri- ods and
covered all peak seasons and several traffic loading scenarios. The
key result that impressed the company's
This content downloaded from 202.43.95.117 on Mon, 18 May 2015
03:33:35 UTCAll use subject to JSTOR Terms and Conditions
-
850 Journal of the Operational Research Society Vol. 53, No,
8
Figure 6 Sample Euclidean routes.
Figure 7 A sub-network for a route.
This content downloaded from 202.43.95.117 on Mon, 18 May 2015
03:33:35 UTCAll use subject to JSTOR Terms and Conditions
-
G loannou et al-Map-Route: a GIS-based decision support system
851
_.!. - - - - - . : . . :: .
.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~--
---
Figure 8 The accurate route.
Figure 9 A zoomed view of the accurate route.
This content downloaded from 202.43.95.117 on Mon, 18 May 2015
03:33:35 UTCAll use subject to JSTOR Terms and Conditions
-
852 Journal of the Operational Research Society Vol. 53, No.
8
management was the possibility of serving all customers even
during the peak seasons under heavy traffic loading conditions.
This was attributed to the fact that optimizing the routes at the
planning stage was indeed better than the intuitive schedules
followed by the drivers. The second aspect of Map-Route that was
in-line with the company's expectations was the simplicity of the
user-interface and the 'power' that the DSS left in the hands of
the users, who were key decision-makers in the daily operations of
the logistics plan. The third positive reaction came from the
drivers, who were asked for a trial period of one week to follow
the schedules produced by Map-Route. They were all able to finish
their delivery routes on-time and served the customers within the
contractual time windows. Thus, even the drivers 'bought-in' the
new application.
Apart from the positive views above, there were some negative
comments by some drivers that apart from full-time employment also
owned their own vehicles that used to 'rent' to the company during
peak seasons. Nevertheless, the obvious savings that the DSS
offered to the company over- came their negative reactions. Table 2
provides a summary of Map-Route implementation details.
It is evident from the results presented in Table 2 that minimal
investment is required to deploy Map-Route, and the cost is
affordable even for SMEs. Furthermore, for the particular
application, Map-Route resulted in effective route planning by
allowing the use of the existing fleet of the company (26
vehicles), even during peak season. The quality of the solution can
be inferred by the significant reduction of both violated time
windows and lost sales; note that these two percentages are
different due to the acceptance of some off-time window deliveries
by several customers. Finally, user training on MapInfo and the
Map-Route components was straightforward and was completed during
the system development (since the two users were involved from the
initial development stages). Unfortunately, we did not have access
to commercial software in order to compare our results.
After the full-scale deployment of Map-Route, several additional
functionalities were requested for implementation.
Table 2 Summary of DSS implementation results
Parameter Status before Map-Route implementation Status after
Map-Route implementation
Required number of vehicles 35 26 Optimized routes No Yes
Violated time windows 20%
-
G loannou et al-Map-Route: a GIS-based decision support system
853
E ?f*Am 21sb it4l ~dow grow$$ YRP LOCA-ROUTES
RREER6AI.IE ~ ~~~IFROMLEFT ITOLEFT IF6RMR GHT1 TOR GHT ILENGTH
IL CROE91I SCODESI L ELTA REA
IERRA ODSoO 2 10; 51' AlIROITO AlIO 11664 1043 IFRATEOR II ~~~
~
~~~ ~~~~~17' 12' 24 12A1100100 AIIOOIOO 11664 10435
IF-ROFANTvIJ1 2: 6: 76: 61100100 A11100100 I11164: 11654:
Z... 1200'i 1EY]6)MOLPIDON I 1, 3, 2 4 38,A112O010_:41165100
11864, 118660 EYM0LPIDVN 1 3' 2' 4~ 38'A11O00_:A 4116I00 11864
11864 IE80FANTVN 10 7 1 12, 6501100100 41162100O 11664..~!,1 11664'
IERIFANTVN 10: 17 lo: i2 58 ;RAl100100 4 l1165100 11854: 11864:
IEROFANTVN 1 11, 21 6 7 . Al100100 AlIIORIOT I 1104. 11664,
FigReA 0 apRotereor on3 the , actua routes. 0 A 000 114:5
145
user-centric philosophy, bringing vehicle routing algorithms
into the hands of planners-schedulers, whose vast experience and
'common sense' can be the determinant success factors for the
GIS-based DSS implementation in real environments. The experiences
from the deployment of Map-Route to an actual case in the Greek
market were presented to demon- strate the phases of the
methodology inherent in the DSS and reveal several open issues that
need to be handled on an exception basis by the users and/or the
knowledge base. Through this case, the flexibility and ease of
adaptation of the DSS were also illustrated.
Via the four-phase approach offered by Map-Route, a user can
easily find a schedule as well as alternative schedules on
intra-city transportation networks for VRPTW. The use of
visualization along with the availability of GIS can help users in
making improved decisions when solving real world routing problems,
which are everyday reality in logistic operations, and become even
more critical due to the expan- sion of third-party logistics.
Thus, developing and deploying effective DSSs is a key prerequisite
for the successful operation of logistics groups. Further
extensions of Map- Route include: (a) integrating modern
meta-heuristics to further improve the quality of the final
solutions; (b) enhan- cing the approach to handle inter-city
networks with addi- tional constraints and route complications; and
(c) integrating the DSS with warehouse management systems (eg,
MANTIS) or Enterprise Resource Planning Systems (eg,
SAP or Oracle Apps) for seamless information technology
applications to distribution problems.
Acknowledgements-The authors would like to thank the two
anonymous referees for their constructive comments and pointers to
archival literature that helped improve the content and the
presentation of the paper.
References
1 Bodin L (1990). Twenty years of routing and scheduling. Opns
Res 38: 571-579.
2 Laporte G (1992). The vehicle routing problem: an overview of
exact and approximate algorithms. Eur J Opl Res 59: 345-358.
3 Gendreau M, Laporte G and Potvin J (1997). Vehicle routing:
modem heuristics. In: Aarts E and Lenstra JK (eds). Local Search in
Combinatorial Optimisation. John Wiley & Sons: Chichester,
UK.
4 Keen PB (1998). Spatial decision support systems for vehicle
routing. Dec Support Syst 22: 65-71.
5 Lapalme G, Rousseau JM, Chapleau S, Cormier M, Cossette P and
Roy S (1992). Geo-route: a geographic information system for
transportation applications. Commun ACM 35: 80-88.
6 Basnet C, Foulds L and Igbaria M (1996). Fleet-Manager: a
microcomputer-based decision support system for vehicle rout- ing.
Dec Support Syst 16: 195-207.
7 Potvin JY, Lapalme G and Rousseau JM (1994). A microcom- puter
assistant for the development of vehicle routing and scheduling
heuristics. Dec Support Syst 12: 41-56.
8 Concialini A and Hasle G (1997). The Greentrip toolkit-
sustainable transportation via intelligent routing. http://
www.cs.strath.ac.uk/research/GreenTrip/ISATA97.doc
This content downloaded from 202.43.95.117 on Mon, 18 May 2015
03:33:35 UTCAll use subject to JSTOR Terms and Conditions
-
854 Journal of the Operational Research Society Vol. 53, No.
8
9 Gambardella LC, Taillard E and Agazzi G (1999). MACS- VRPTW: a
multiple ant colony system for vehicle routing problems with time
windows. In: Come D, Dorigo M and Glover F (eds). New Ideas in
Optimization. McGraw-Hill: London, UK, pp 63-76.
10 Wahle J, Annen 0, Schuster C, Neubert L and Schreckendberg M
(2001). A dynamic route guidance system based on real traffic data.
Eur J Opl Res 131-132: 74-80.
11 Savelsberg M and Sol M (1998). DRIVE: dynamic routing of
independent vehicles. Opns Res 46: 474-490.
12 Campbell JF and Langevin A (2000). Arc routing applications
for roadway snow and ice control. In: Dror M (ed). Arc Routing:
Theory, Solutions and Applications. Kluwer: Amsterdam.
13 Golden BL and Wasil EA (1987). Computerized vehicle routing
in the soft drink industry. Opns Res 35: 6-17.
14 Ioannou G, Kritikos M and Prastacos G (2001). A greedy look-
ahead heuristic for the vehicle routing problem with time windows.
J Opl Res Soc 52: 523-537.
15 Solomon MM (1987). Algorithms for the vehicle routing and
scheduling problems with time window constraints. Opns Res 35:
254-265.
16 MapInfo (2000). MapInfo Professional: Users Guide. MapInfo
Corporation: Troy, NY.
17 Taylor MA (1990). Knowledge-based systems for transport
network analysis: a fifth generation perspective on transport
network problems. Transport Res A 24: 3-14.
18 Cormen TH, Leiserson CE and Rivest RL (1992). Introduction to
Algorithms. MIT Press: Cambridge, MA.
19 Kritikos MN (1997). Mathematical models and algorithmic
approaches for transportation problems using Geographical
Information Systems. PhD dissertation, Athens University of
Economics and Business.
Received September 2001; accepted January 2002 after one
revision
This content downloaded from 202.43.95.117 on Mon, 18 May 2015
03:33:35 UTCAll use subject to JSTOR Terms and Conditions
Article Contentsp. [842]p. 843p. 844p. 845p. 846p. 847p. 848p.
849p. 850p. 851p. 852p. 853p. 854
Issue Table of ContentsJournal of the Operational Research
Society, Vol. 53, No. 8, Aug., 2002Front MatterCase-Oriented
PapersRevenue Impacts of Fare Input and Demand Forecast Accuracy in
Airline Yield Management [pp. 811 - 821]A Multidimensional Knapsack
Model for Asset-Backed Securitization [pp. 822 - 832]Prototype
Fleet Optimization Model [pp. 833 - 841]Map-Route: A GIS-Based
Decision Support System for Intra-City Vehicle Routing with Time
Windows [pp. 842 - 854]
Theoretical PapersDynamic Demand Lot-Sizing Rules for
Incremental Quantity Discounts [pp. 855 - 863]The Wafer Probing
Scheduling Problem (WPSP) [pp. 864 - 874]Customer Knowledge
Management [pp. 875 - 884]Supply Chain Modelling and Its Analytical
Evaluation [pp. 885 - 894]Comparing an ACO Algorithm with Other
Heuristics for the Single Machine Scheduling Problem with
Sequence-Dependent Setup Times [pp. 895 - 906]A Comparison of the
Performance of Artificial Intelligence Techniques for Optimizing
the Number of Kanbans [pp. 907 - 914]
Technical NotesOn the Economic Order Quantity under Conditions
of Permissible Delay in Payments [pp. 915 - 918]A Generalised
Life-Expectancy Model for a Population [pp. 919 - 921]A
Manufacturer's Optimal Quantity Discount Strategy and Return Policy
Through Game-Theoretic Approach [pp. 922 - 926]
ViewpointsTechnical Note on Balanced Solutions in Goal
Programming, Compromise Programming and Reference Point Method [pp.
927 - 929]Reply to the Comments of Ganjavi et al [pp. 929 -
930]Response to Reply to the Comments of Ganjavi, Aouni and Wang
2002 [pp. 930 - 931]Final Reply to the Comments of Professors
Ganjavi et al [p. 931]
Back Matter