ISSN 1748-7595 (Online)
Kent
Business
School
Working Paper Series
Location-routing: Issues, Models
and Methods
Said Salhi
Gabor N agy
Kent Business School
Working Paper N o.124
M ay 2 0 0 6
Location-routing: Issues, models and methods*
Gábor Nagy and Saïd Salhi
Centre for Heuristic Optimisation, Kent Business School, The University of Kent,
Canterbury, Kent, CT2 7PE, U.K.
{G.Nagy,S.Salhi}@kent.ac.uk
Abstract: This paper is a survey of location-routing: a relatively new branch of
locational analysis that takes into account vehicle routing aspects. We propose a
classification scheme and look at a number of problem variants. Both exact and
heuristic algorithms are investigated. Finally, some suggestions for future
research are presented.
Keywords: combinatorial optimisation, heuristics, location, logistics, routing
1. Introduction
The aim of this paper is to survey the state of the art in location-routing. The location-routing
problem (LRP) is a research area within locational analysis, with the distinguishing property of paying
special attention to underlying issues of vehicle routing. Although there are a large number of surveys
on various aspects of location theory (see EWGLA (2003)), the LRP received little attention, with
previous surveys by Balakrishnan, Ward and Wong (1987), Laporte (1988, 1989), Berman, Jaillet and
Simchi-Levi (1995) and Min, Jayaraman and Srivastava (1998). Furthermore, research has moved on
considerably since these works – about a third of the papers we report on appeared since the last
review. Thus, we felt that a new state-of-the-art survey was timely and desirable.
We intend to make this paper both a comprehensive review of location-routing and an accessible
introduction for those working in other areas of location theory. We particularly hope it will be a
useful guide to doctoral students who wish to begin their research career in this area. We wish to
include all the journal papers on location-routing and make references to related research areas. For
those who are new to the wider research field of location, an extensive list of introductory textbooks
and survey papers is given in EWGLA (2003). For vehicle routing, we can recommend Christofides
et al. (1979), Laporte (2000) and Toth and Vigo (2002a, 2002b).
1.1. Definition
The phrase “location-routing problem” is misleading, as location-routing is not a single welldefined problem like the Weber problem or the travelling salesman problem. It can be thought of as a
set of problems within location theory. However, we prefer to think of the LRP as an approach to
modelling and solving locational problems. Thus, we define location-routing, following Bruns
(1998), as “location planning with tour planning aspects taken into account” (“Standortplanung unter
*
This paper is accepted for publication by the European Journal of Operational Research.
1
Berücksichtigung von Tourenplanungsaspekten”). This is in line with Balakrishnan et al. (1987,
p.56.), who observe that “location/routing problems are essentially strategic decisions concerning …
facility location.” Our definition stems from a hierarchical viewpoint, whereby our aim is to solve a
facility location problem (the “master problem”), but in order to achieve this we simultaneously need
to solve a vehicle routing problem (the “subproblem”). This also implies an integrated solution
approach, i.e. we do not classify as belonging to the LRP an approach that deals with both location
and routing aspects of a problem but does not address their inter-relation. Another important
characteristic of our definition is the requirement for the existence of tour planning, i.e. the existence
of multiple stops on routes. This occurs if customer demands are less than a full truckload. Other
authors coin different definitions of the LRP, some including a wider range of problem versions than
we do.
1.2. Location, routing and location-routing
It is well known that facility location and vehicle routing are interrelated areas. Maranzana (1964,
p.261.) points out that “the location of factories, warehouses and supply points in general … is often
influenced by transport costs.” (Some consider this paper as the first publication on the LRP,
although strictly speaking it incorporates shortest-path, rather than vehicle-routing, problems into a
locational problem.) Further to the above, Rand (1976, p.248.) observes that “many practitioners are
aware of the danger of suboptimizing by separating depot location and vehicle routing.” However,
both academics and practitioners often ignore this interrelation, and solve locational problems without
paying attention to underlying routing considerations. We list below three possible reasons for this.
(1) There are many practical situations when locational problems do not have a routing aspect. In
these cases, the location-routing approach is clearly not an appropriate one.
(2) Some researchers object to location-routing on the basis of a perceived inconsistency. They point
out that location is a strategic, while routing is a tactical problem: routes can be re-calculated and
re-drawn frequently (even daily), depot locations are normally for a much longer period. Thus,
they claim that it is inappropriate to combine location and routing in the same planning framework
due to their different planning horizons. This criticism led the authors to investigate this issue: it
was found that the use of location-routing could decrease costs over a long planning horizon,
within which routes are allowed to change. (For a more detailed discussion on this issue, see Salhi
and Nagy (1999)).
(3) The LRP is conceptually more difficult than the classical location problem. Berman et al. (1995,
p.431.) observe that in the LRP, “the facility … must be “central” relative to the ensemble of the
demand points, as ordered by the (yet unknown) tour through all of them. By contrast, in the
classical problems the facility … must be located by considering distances to individual demand
points, thus making the problem more tractable.” This may have also contributed to slow progress
on the LRP.
Location-routing problems are clearly related to both the classical location problem and the vehicle
routing problem. In fact, both of the latter problems can be viewed as special cases of the LRP. If we
require all customers to be directly linked to a depot, the LRP becomes a standard location problem.
If, on the other hand, we fix the depot locations, the LRP reduces to a VRP. From a practical
viewpoint, location-routing forms part of distribution management, while from a mathematical point
of view, it can usually be modelled as a combinatorial optimisation problem. We note that this is an
NP-hard problem, as it encompasses two NP-hard problems (facility location and vehicle routing).
Since a number of problem versions exist, we cannot reproduce all the formulations here. In the first
instance, the reader is referred to Laporte (1988) for an excellent review of various formulations.
Table 1 presents a summary of formulations for a variety of LRP versions developed since the
publication of the above review.
2
Problem type
Stochastic LRP
Dynamic LRP
Hamiltonian p-median
Road-train routing
Vehicle routing–allocation (VRAP)
Many-to-many LRP
Eulerian location
LRP with mixed fleet
Location-routing-inventory
Plant cycle location
Many-to-many LRP
Multi-level location-routing-inventory
Deterministic LRP
VRAP (median cycle problem)
LRP with non-linear costs
Planar LRP (single-depot)
Restricted VRAP
Planar LRP (multi-depot)
Paper
Laporte et al. (1989)
Laporte and Dejax (1989)
Branco and Coelho (1990)
Semet (1995)
Nascimento and Beasley (1996)
Nagy and Salhi (1998)
Ghiani and Laporte (1999)
Wu et al. (2002)
Liu and Lee (2003)
Labbé et al. (2004)
Wasner and Zäpfel (2004)
Ambrosino and Scutellà (2005)
Albareda-Sambola et al. (2005)
Labbé et al. (2005)
Melechovský et al. (2005)
Schwardt and Dethloff (2005)
Gunnarsson et al. (2006)
Salhi and Nagy (2007)
Section
5.2
5.3
4.2
6.4
6.3
6.2
3
4.3
5.2
3
6.2
6.4
4.4
6.3
4.4
4.2
6.3
4.3
Table 1. ILP formulations for various LRP problems
Finally, we wish to point out that an integrated approach to solving distribution management
problems is not restricted only to location-routing. Integrated approaches are becoming more popular
and both the location and the routing problems have been studied in conjunction with other logistical
problems. Although such problems are not within the remit of this article, readers interested in
combined logistics problems in general may find the following summary useful. Location aspects are
present in:
(a) the queueing–location problem (reviewed by Berman and Krass (2001) and Boffey et al. (2007)),
(b) the location–assignment problem (Maze and Khasnabis (1985)),
(c) the location–capacity-acquisition problem (Verter and Dincer (1995)),
(d) the location–network-design problem (Melkote and Daskin (2001), Lee et al. (2003)),
(e) the inventory–location problem (Daskin et al. (2002), Shen et al. (2003), Drezner et al. (2003)) and
(f) the location–scheduling problem (Hennes and Hamacher (2006)).
Routing forms part of the following combined logistics problems:
(g) the inventory–routing problem (reviewed by Baita et al. (1998) and Moin and Salhi (2007)),
(h) the routing–scheduling problem (Metters (1996), Averbakh and Berman (1999)) and
(i) the routing–packing problem (Türkay and Emel (2007), Ferrer, Nagy and Wassan (2007)).
However, very few authors investigate integrating location-routing with other aspects of distribution
management. Murty and Djang (1999) (see 6.3) look at a combined location–routing–scheduling
problem. Both Liu and Lee (2003) (see 5.2) and Ambrosino and Scutellà (2005) (see 6.4) include
inventory aspects in the LRP.
1.3. Applications of location-routing
Operational Research is primarily an applications-oriented discipline. Therefore, we thought it
important to highlight practical applications of location-routing. Table 2 summarises the main
characteristics of papers describing practical applications, giving reference to the section where they
will be discussed in detail. It also shows the size of the largest instances solved, in terms of the
number of potential facilities and number of customers.
3
Paper
Watson-Gandy and Dohrn (1973)
Bednar and Strohmeier (1979)
Or and Pierskalla (1979)
Jacobsen and Madsen (1980)
Nambiar et al. (1981)
Perl and Daskin (1984, 1985)
Labbé and Laporte (1986)
Nambiar et al. (1989)
Semet and Taillard (1993)
Kulcar (1996)
Murty and Djang (1999)
Bruns et al. (2000)
Chan et al. (2001)
Lin et al. (2002)
Lee et al. (2003)
Wasner and Zäpfel (2004)
Billionnet et al. (2005)
Gunnarsson et al. (2006)
Lischak and Triesch (2007)
Section
4.4
4.2
4.4
6.4
4.1/4.4
4.3
6.3
5.3
6.4
6.1
6.3
6.2
5.3
4.4
3
6.2
3
6.3
6.2
Application area
Food and drink distribution
Consumer goods distribution
Blood bank location
Newspaper distribution
Rubber plant location
Goods distribution
Postbox location
Rubber plant location
Grocery distribution
Waste collection
Military equipment location
Parcel delivery
Medical evacuation
Bill delivery
Optical network design
Parcel delivery
Telecom network design
Shipping industry
Parcel delivery
Country
United Kingdom
Austria
United States
Denmark
Malaysia
United States
Belgium
Malaysia
Switzerland
Belgium
United States
Switzerland
United States
Hong Kong
Korea
Austria
France
Europe
Poland
Facilities
Customers
40
3
3
42
15
4
300
50
117
4510
300
318
Not given
Not given
10
9
13
29
200
9
4
50
10
6
24
22
47
90
260
331
3200
52
27
50
2042
70
300
750
Table 2. A summary of LRP applications
We can see that practical problems with hundreds of possible depot locations and thousands of
customers can be solved. These papers show an enormous variety. Although most of them focus on
distribution of consumer goods or parcels, there are also some applications in health, military and
communications. Operational Research is all too often applied only in the affluent countries of
Western Europe and North America, thus it is pleasing to see that LRP has also been applied in
developing countries. We also note that application-oriented papers account for about a fifth of the
LRP literature. The above observations show that LRP is really applicable in practice and is not just a
purely academic construct.
1.4. Selection criteria and organisation of the paper
Even though location-routing is a fairly small research area, some selection was needed to keep the
paper within reasonable lengths. Thus, we decided to omit working papers, conference publications
and book chapters from our review and focus only on journal articles. (However, we included a few
papers that are currently in the “publication pipeline”.) Nor do we discuss dissertations within this
article, however we do feel that it is encouraging to see a steady stream of theses devoted to this topic,
such as Perl (1983), Reece (1985), Srivastava (1986), Salhi (1987), Simchi-Levi (1987), Nagy (1996),
Berger (1997), Bruns (1998), Lischak (2001), Albareda-Sambola (2003), Çetiner (2003), Souid
(2003), Barreto (2004) and Tham (2005). (The theses of Melkote (1996) and Rodríguez-Martín
(2000) address related topics.) To show the development of ideas, we usually follow a chronological
order, but sometimes deviate from it to group similar problem versions or methods together. (Note
that the chronological order relates to publication date, unless a paper published earlier is a follow-up
to one published later). We have aimed to be as comprehensive as possible, but we apologise if we
have inadvertently omitted any articles.
Some papers and areas will be reviewed in more details than others. To some extent, this is
determined by our personal taste and judgement about their importance. We wish to focus on the
models and variants rather than technical improvements and results. Areas that are related to but not
strictly within LRP are only reviewed briefly. Furthermore, our reviews of exact methods and
stochastic problems are written fairly concisely, so as to minimise overlap with the reviews of Laporte
4
(1989) and Berman, Jaillet and Simchi-Levi (1995), who provide very extensive descriptions of these
topics.
The remainder of the paper is structured as follows. The next section presents our classification
scheme. Sections 3 and 4 look respectively at exact and heuristic approaches for static deterministic
problems. Section 5 is devoted to stochastic and dynamic problems. Problems with non-standard
hierarchical structures are investigated in section 6. Finally, we list some suggestions for future work
in section 7. Each of sections 3 to 6 contains a summary table. These list the most recent major
development for each problem version and solution method. Hence, these articles date mainly from
the last decade, although some older papers are included for the sake of completeness. These tables
also show the problem type and solution method, together with the size of the largest instance solved
(in terms of the number of potential facilities and number of customers).
2. Classification and reader’s guide
As it is possible in theory to develop a location-routing version of just about all location problems,
classifying location-routing problems is at least as difficult a task as that of classifying location
problems, with added complexity provided by the variability in the underlying vehicle routing
problems. Clearly, there are a number of different ways of classification, involving some element of
arbitrary choice. We look at eight aspects of the problem structure, pertaining to the location of
facilities, the pattern of vehicle routes or to the entirety of the problem. Furthermore, the type of
solution method is also used as a classification criterion. Our aim in choosing these criteria was to
identify various strands of research. We found that the resulting groupings of papers correspond well
to the – sadly, somewhat disjoint – efforts of the LRP research community. However, we are at pains
to point out connections between various areas and methods, to provide a more integrated view of the
LRP. We will also take advantage of this section to describe various aspects of the problem.
Our survey of the literature is structured according to this classification scheme. Thus, this section
also serves as a reader’s guide, explaining the allocation of papers to sections. In particular, readers
interested only in certain aspects of, or approaches for, the LRP are recommended to use this section
as a guide of quickly finding papers pertaining to their fields of interest.
We begin our classification by looking at four key aspects of location-routing problems. These
will form the basis of the structure of the paper.
(a) Hierarchical structure. The structure of most location-routing problems consist of facilities
servicing a number of customers, these are connected to their depot by means of vehicle tours. No
routes connect facilities to each other. However, there is a body of literature that deviates from
this structure. Some of these works represent quite complex extensions to the LRP; some others
may not even be considered to be part of the LRP. Hence, the above standard hierarchical
structure will be assumed through sections 3 to 5, and all deviations from it discussed separately in
section 6.
(b) Type of input data: this may be deterministic or stochastic. There is a larger body of literature on
the deterministic case. We note that all stochastic papers consider customer demand as the only
stochastic variable. Stochastic papers will be the subject of section 5.
(c) Planning period: this may be single-period or multi-period. Problems with single or multiple
periods are known respectively as static or dynamic. The vast majority of LRP papers investigate
the static case. Dynamic papers will be described in section 5, together with stochastic papers.
(d) Solution method: this may be exact or heuristic. There are more papers using heuristic methods,
but exact methods are often very successful for special cases of the LRP. Sections 3 and 4 will
look at exact and heuristic methods, respectively, for static deterministic problems. In sections 5
5
and 6, exact and heuristic approaches are discussed together. In the context of heuristic methods,
we wish to point out a peculiarity of this research field: research is so fragmented that only four
papers furnish computational comparisons to their peers. Thus, our comparative analysis of
heuristics will be more often qualitative than quantitative.
It is clearly not possible to describe together all combinations of papers that are similar in one
respect or another. Neither do we wish to follow a complete taxonomy, as this would create a very
large number of groupings, each containing only a few papers and this would not allow us to show the
logical development of ideas. Thus, the remainder of this section will allow the reader to find papers
according to classification criteria further to the ones discussed above. (Numbers in brackets after
papers refer to the section or subsection where that paper is described.) We also aim to break down
the rigidity of the structure by cross-referring between sections as appropriate.
(e) Objective function. The usual objective for location-routing problems is that of overall cost
minimisation, where costs can be divided into depot costs and vehicle costs. There are only a few
papers where a different objective prevails or consider multiple objectives. These are: Averbakh
and Berman (1994, 2002) [3], Averbakh and Berman (1995), Averbakh et al. (1994) and Jamil et
al. (1994) [5.1]. Furthermore, most of the literature on the related problem of transportationlocation (see 6.1) is multi-objective.
(f) Solution space. This can be discrete, network or continuous. Most of the LRP literature deals
with discrete location. However, many works on the round-trip location problem (see section 3)
and the travelling salesman location problem (see 5.1 and also Simchi-Levi (1991) in 5.2) are
restricted to path or tree networks. Planar location is also often considered for the above problem
variants, but apart from these cases only Schwardt and Dethloff (2005) [4.2] and Salhi and Nagy
(2007) [4.3] deal with continuous problems.
(g) Number of depots. This may be single or multiple. Most papers on the LRP deal with multiple
depots, except Laporte and Nobert (1981), Averbakh and Berman (1994, 2002) [3], Simchi-Levi
(1991) [5.2] and Schwardt and Dethloff (2005) [4.2] who restrict themselves to single depots.
However, some special cases are solved only for one depot, such as the travelling salesman
location problem [5.1] and most round-trip location problems [3]. Furthermore, when multiple
depots are considered, it is generally assumed that the number of depots is not given in advance,
the exceptions to this being Branco and Coelho (1990) [4.2] and Salhi and Nagy (2007) [4.3].
(h) Number and types of vehicles. For most location-routing problems, the number of vehicles is not
fixed in advance and a homogeneous fleet is assumed. However, a heterogeneous fleet is
considered by Bookbinder and Reece (1988), Salhi and Fraser (1996), Wu et al. (2002) [4.3],
Ambrosino and Scutellà (2005) [6.4] and Gunnarsson et al. (2006) [6.3]. Laporte and Nobert
(1981) and Averbakh and Berman (2002) [3] investigate problems when the number of vehicles is
given in advance. Furthermore, Laporte, Nobert and Pelletier (1983) [3] and Branco and Coelho
(1990) [4.2] look at the special case of exactly one vehicle per depot. Of necessity, travelling
salesman location problems (see 5.1) also have this structure.
(i) Route structure. The usual structure of vehicle routes in an LRP is to start out from a depot,
traverse through a number of customer nodes, delivering goods at each customer and finally return
to the same depot. For most location-routing problems, this structure holds true, but we note here
the following exceptions. Vehicles may traverse given edges rather than nodes (known as arc
routing), see Levy and Bodin (1989) [4.2] and Ghiani and Laporte (1999) [3]. Vehicles may be
allowed multiple trips, see Lin, Chow and Chen (2002) [4.4]. Vehicle routes may contain both
deliveries and pickups, see Mosheiov (1995) [5.1], the round-trip location problem [3] and the
many-to-many LRP [6.2].
6
3. Exact solution methods for deterministic problems
The first deterministic location-routing type problem to be solved to optimality was the round-trip
location problem, a special case with the following route structure. Vehicles start out from a depot,
visit a customer to pick up some load, deliver it to another customer and then return to the depot. This
problem occurs frequently in practice (e.g. courier service). It was introduced by Chan and Hearn
(1977), who assume that customers are located on a plane with rectilinear distances. The special
properties of the problem make it possible to find efficiently the minimum value of the objective
function. Then, by substituting this value into a linear programme, the optimal solution can be found.
Chan and Francis (1976) solve the case when customers are located on a tree graph using a similar
procedure. The algorithm of Drezner and Wesolowsky (1982) is based on the numerical solution of
differential equations. Further solution algorithms of this type are also presented by Drezner (1982).
The last two papers both solve planar problems with a variety of distance norms. While all of the
above papers concern the location of a single facility, Kolen (1985) generalises the problem to
locating several facilities. An underlying tree network is assumed and a constructive optimal
algorithm that iteratively partitions the tree is presented. Round-trip location problems can be solved
to optimality for very large problems within reasonable computing times: Drezner (1982) solves
problems with up to 5000 pairs of demand points.
Exact methods for more general LRPs are usually based on a mathematical programming
formulation. They often involve the relaxation and reintroduction of constraints such as: (a) subtour
elimination (all vehicle tours must contain a depot), (b) chain barring (routes are not allowed to
connect one depot to another) and (c) integrality (certain variables must be integer – usually binary
integer). The following four papers all begin with relaxing these constraints.
The first exact algorithm for the general LRP is by Laporte and Nobert (1981); in this paper a
single depot is to be selected and a fixed number of vehicles is to be used. A branch-and-bound
algorithm is used. The authors note that the optimal depot location rarely coincides with the node
closest to the centre of gravity.
Laporte, Nobert and Pelletier (1983) consider locating several depots, with or without depot fixed
costs and with or without an upper limit on the number of depots. For the special case of only one
vehicle per depot, it was found to be more efficient to first reintroduce subtour elimination constraints
(there would be no chain barring constraints) and then use Gomory cuts to achieve integrality.
Otherwise, the authors recommend using Gomory cuts first and then reintroducing subtour and chain
barring constraints. On the other hand, the method of Laporte, Nobert and Arpin (1986) applies a
branching procedure where subtour elimination and chain barring constraints are reintroduced. (A
similar approach is adopted by Laporte, Louveaux and Mercure (1989) for a stochastic LRP, see 5.2.)
Laporte, Nobert and Taillefer (1988) use a graph transformation to reformulate the LRP into a
travelling salesman type problem. They apply a branch-and-bound algorithm, where in the search
tree, each subproblem is a constrained assignment problem and can thus be solved efficiently. This
approach is extended to the dynamic LRP by Laporte and Dejax (1989), see 5.3.
The delivery man problem is a version of the TSP where the objective is to minimise the total
waiting time of all customers. The sales-delivery man problem combines the above objective with the
usual TSP-objective of minimising total tour length. Averbakh and Berman (1994) consider the
problem of finding the home base of one or several delivery or sales-delivery men. Polynomial-time
algorithms are presented for location on a path.
Ghiani and Laporte (1999) investigate the Eulerian location problem introduced by Levy and
Bodin (1989) (see 4.2). In this problem the routes, instead of consisting of customer nodes, require
the vehicles to traverse given edges (i.e. the underlying routing problem is the arc routing problem).
In this paper, both maximum capacity and route length constraints are absent. The solution method
begins by transforming the problem to either the rural postman problem or a relaxation thereof. The
solution is then found by constraint relaxation and branch-and-cut.
7
Averbakh and Berman (2002) introduce the minmax p-travelling salesmen location problem,
where the objective is to minimise the length of the longest vehicle tour. In their problem, customers
are at the vertices of a tree, a single depot must be located on a vertex or an edge of the tree and the
number of vehicles is set in advance. An optimal solution is found by reducing the problem to the
minimal γ-dividing set problem.
The plant-cycle location problem was introduced by Billionnet, Elloumi and Grouz-Djerbi (2005).
They consider the problem of simultaneously locating radio-communication stations and designing
rings that connect radio antennae to such stations. This problem only differs from the LRP in that
instead of vehicle routes, rings of communications links are constructed. There is a maximum limit
on the number of antennae per ring, corresponding to “vehicle capacity”. Furthermore, the capacity of
the stations can be chosen and their costs depend on the chosen size. Optimal solution is found by
commercial software. The above problem is further studied by Labbé, Rodríguez-Martin and SalazarGonzález (2004). Integrality and connectivity constraints are relaxed but a number of valid
inequalities are added. An initial solution is found by a simple heuristic and then a branch-and-cut
method is used, adding violated inequalities and improving on the routes in each step using the same
heuristic. We note that, as no actual vehicle exists in this problem, it could also be viewed as a
combined location and network design problem. (For networks that are more complex than a set of
circuits the location–network design problem is no longer an LRP and thus beyond the scope of this
review; the reader is referred to Melkote and Daskin (2001) and to Lee et al. (2003).)
In summary, exact methods provide significant insights into problems, but due to the complexity
of location-routing they can only tackle relatively small instances. General location-routing instances
with up to 40 potential depot locations or 80 customers have been solved to optimality (Laporte et al.
(1983, 1988)). However, exact methods can be very successful for solving special cases of the LRP,
in particular the round-trip location problem. Table 3 summarises, for each problem type and solution
method, the most recent paper and the size of the largest problem solved (if available).
Problem type
General deterministic LRP
Round-trip location
Eulerian location
Minmax TS location
Plant cycle location
Solution method
Cutting planes
Branch-and-bound
Numerical optimisation
Branch-and-cut
Graph theoretical
Branch-and-cut
Paper
Laporte et al. (1983)
Laporte et al. (1988)
Drezner (1982)
Ghiani and Laporte (1999)
Averbakh and Berman (2002)
Labbé et al. (2004)
Facilities
40
3
1
50
1
30
Customers
40
80
10000
200
Not given
120
Table 3. Summary of recent papers on exact methods for deterministic problems.
4. Heuristic solution methods for deterministic problems
Due to the relatively large number of papers devoted to this area, we decided to further classify
them according to the solution method employed. We describe our classification scheme in the next
subsection. The subsequent three subsections look at clustering-based, iterative and hierarchical
heuristics, respectively.
4.1. A classification and an overview of LRP heuristics
We decided to base our classification on how the solution algorithms model the relationship
between the locational and the routing subproblems.
8
Sequential methods first solve the locational problem by minimising the sum of depot-to-customer
distances (also known as radial distances) and then solve the routing problem based on the depot
locations found. (A more sophisticated approach is to use route length estimation formulae instead of
radial distances, as advocated by Webb (1968) and Christofides and Eilon (1969) and first used by
Watson-Gandy and Dohrn (1973). For more information on this topic, see for example Daganzo
(2005).) The sequential solution concept does not allow for feedback from the routing phase to the
locational phase. Balakrishnan et al. (1987, p.37.) point out that “the sequential solution of a classical
facility location and a vehicle routing model can … lead to a suboptimal design for the distribution
system.” This observation was later supported by the empirical investigations of Salhi and Rand
(1989) and Salhi and Nagy (1999) for the static and the dynamic cases respectively. However,
Srivastava and Benton (1990) observe that sequential methods are capable in some cases of providing
good quality solutions. We also note that solving both locational and routing subproblems to
optimality in sequence would still be a heuristic, as it cannot guarantee an optimal solution to the
combined problem. We do not classify sequential methods as part of the LRP approach and hence
exclude them from our review, except for a few that pertain to interesting special cases. However, we
note that sequential methods are useful for benchmarking other heuristics when no other evaluation
measure is available.
Clustering-based methods begin by partitioning the customer set into clusters: one cluster per
potential depot or one per vehicle route. Then, they may proceed in two different ways:
(a) locating a depot in each cluster and then solving a VRP (or TSP) for each cluster;
(b) solving a TSP for each cluster and then locating the depots.
In some respect, they resemble sequential methods as no feedback takes place. However, clustering is
based on some “skeleton” of a routing plan (such as a minimal spanning tree of all customers), so this
is a better attempt at integrating locational and routing decisions.
Iterative LRP heuristics decompose the problem into its two constituent subproblems. Then, the
methods iteratively solve the subproblems, feeding information from one phase to the other. Clearly,
the crux of the problem here is just how information can be compressed from one phase and fed into
the other.
Hierarchical heuristics are motivated by the following observations. While iterative methods are
clearly an improvement on sequential methods, they “inherit” some of their drawbacks. They let the
locational algorithm run until the end and then re-start it taking into account new routing information.
Thus, if the routing information is not utilised well in the location phase, the method may go astray.
We can also object to iterative methods from a modelling point of view. They treat the two
constituent components as if they were on the same footing. This is not in line with our view of a
hierarchical structure, with location as the main problem and routing as a subordinate problem. We
conceive heuristic algorithms where the main algorithm is devoted to solving the location problem
and refers in each step to a subroutine that solves the routing problem. We believe such hierarchical
methods may provide a better model of the real situation and are also likely to give better solutions.
These methods sometimes rely on route length estimation.
A summary of the most important recent heuristic papers is given in Table 4.
Problem type
General deterministic LRP
Plant cycle location
Planar LRP
Solution method
Clustering-based
Iterative
Hierarchical
Hierarchical
Hierarchical
Clustering-based
Iterative
Paper
Barreto et al. (2007)
Salhi and Fraser (1996)
Nagy and Salhi (1996b)
Albareda-Sambola et al. (2005)
Melechovský et al. (2005)
Billionet et al. (2005)
Salhi and Nagy (2007)
Facilities
Customers
15
199
400
10
20
6
Infinite
Table 4. Summary of recent papers on heuristic methods for deterministic problems.
9
318
199
400
30
240
70
199
4.2. Clustering-based methods
Bednar and Strohmeier (1979) partition customers into one, two or three clusters, on the basis of
Euclidean depot-to-customer distances weighted with customer demands. This is used to see if
locations proposed by their client are reasonable. Then, VRPs with the proposed locations are solved
using a savings method.
Nambiar, Gelders and Van Wassenhove (1981) consider the case of locating a single depot. First,
customers are clustered according to the capacity and the maximum distance constraints of the
vehicles. This is based on the idea that every cluster should be reachable from every potential depot
location by a feasible vehicle tour. Then, for each potential depot and each cluster a TSP is solved.
The depot with the least cost is selected and a VRP is solved to improve the TSP routes. This method
is clearly applicable to only a small number of potential depots. (The authors also look at the case of
multiple depots, see 4.3.)
Branco and Coelho (1990) look at a special case of the LRP, called the Hamiltonian p-median
problem, where exactly p facilities must be located and each facility has exactly one route. This is
achieved by partitioning the customer set into p Hamiltonian circuits.
The “cluster-routing” heuristic of Srivastava and Benton (1990) and Srivastava (1993) considers in
turn p = m, m–1, …, 2, 1 depots being open, where m is the number of potential depots. For each
value of p, it partitions the customer set into p clusters, based on the minimal spanning tree. For each
cluster, the depot is located at the site nearest to the cluster centre. Then, the resulting VRPs are
solved using the sweep method. The methods compare favourably against a sequential method.
Min (1996) considers the problem of locating consolidation terminals. In this problem, goods
from several supply sources are aggregated at terminals before being sent to customers. This problem
is somewhat more complicated than the basic LRP in that a number of supply points are present and
the allocation of both customers and suppliers to terminals needs to be found. Customers are clustered
according to vehicle capacity and the centroid of each cluster is then used in locating the terminals.
Billionnet, Elloumi and Grouz-Djerbi (2005) solve the plant cycle location problem, which is
structurally nearly equivalent to the LRP (see section 3). First, a minimal spanning forest problem is
solved heuristically. This determines which plants should be open and also finds the allocation of the
customers to plants. The resulting vehicle routing problems are then solved using the savings method.
The authors also derive an improved lower bound.
The heuristic proposed by Barreto et al. (2007) begins by clustering customers according to the
capacity of the vehicles. Then, for each cluster, a TSP is solved – optimally for small clusters and
heuristically, using the savings method and 3-opt, for large clusters. Finally, depot locations are found
by treating each tour as a single customer.
We mention here two papers that are similar to clustering methods in that a one-pass procedure
with no feedback is used but information about route structures is nonetheless incorporated in the
location algorithm. Levy and Bodin (1989) introduce the Eulerian location problem, where the
vehicles are required to traverse given edges (rather than nodes). Depots are located based on the
partition network (an extension of the underlying road network) and an attractiveness measure relating
to the number and weight of the arcs incident to potential depot nodes. Routes are found by a rural
postman problem algorithm. Schwardt and Dethloff (2005) solve a planar LRP with a single depot
and model the relationship of depot and customers by interconnected neuron rings. As the number of
routes must be fixed for this approach (each route corresponds to a neuron ring), the method is re-run
with varying number of vehicles. The resulting depot location is then improved upon by solving a
Weber problem with the end-points of the tours as the fixed points. As the neural network algorithm
is a stochastic procedure, the authors repeat their method several times to find the best solution. The
results compare favourably to those found by a sequential procedure. Finally, we note that clusteringbased methods are also used to solve a dynamic LRP (see 5.3) by Chan, Carter and Burnes (2001) and
a road-traing routing problem (see 6.4.) by Semet (1995). Such methods are also used to generate
possible starting points for other heuristics.
10
4.3. Iterative methods
Perl and Daskin (1984, 1985) first introduced the concept of iterating between locational and
routing phases. The locational phase is formulated as an ILP and solved to optimality using implicit
enumeration. It minimises the sum of distances between depots and the “end-points” of routes found
in the routing phase. The routing phase uses a savings-type heuristic generalised for multiple depots.
The procedure terminates when in either phase the cost improvement is zero or negligible.
This method was later improved by several authors. Hansen et al. (1994) solve both phases
heuristically, this allows more time for routing calculations and thus leads to better solution quality.
Salhi and Fraser (1996) consider not just the end-points of tours as input to the locational phase but
investigate all pairs of customers. They also include the length of tours in calculating the variable
costs of the location model. Furthermore, these variable costs are adjusted to take into account a
heterogeneous fleet. The locational phase is built on the moves drop and shift and the routing phase is
based on a multi-level fleet mix heuristic. Reasonable improvement is found when compared against
a sequential method. The method of Wu, Low and Bai (2002) is similar to the above, except that both
phases rely on a combined tabu search and simulated annealing framework, but with a simpler
neighbourhood structure. On some problem sets, this procedure outperforms Perl and Daskin (1984)
and Hansen et al. (1994). However, no comparison with Salhi and Fraser (1996) is given.
The “end-point” concept is also used by Salhi and Nagy (2007) who solve a planar LRP with a
fixed number of depots. In the locational phase, a Weber problem is solved for each depot with the
end-points of the tours found in the routing phase as customers. The routing phase consists of a multidepot savings-based heuristic with several improvement routines. The two phases are repeated until
no significant improvement is found. The method is shown to improve on the results of a sequential
approach and produces results similar to Schwardt and Dethloff (2005).
A different iterative framework is used by Bookbinder and Reece (1988) who apply Benders
decomposition to split the LRP into location-allocation and routing subproblems. (The latter is
extended to allow for heterogeneous fleet.) These are solved to optimality in an iterative framework.
An advantage of this approach is that it produces upper and lower bounds in each iteration.
We note that an iterative method is also used by Labbé and Laporte (1986) to solve the vehicle
routing-allocation problem (see 6.3).
4.4. Hierarchical methods
Nambiar, Gelders and Van Wassenhove (1981) present a method that uses the result of their
single-depot clustering heuristic as the starting point. Then, they consider in turn p = 1, 2, …, m
depots being open. For each value of p, they reformulate the LRP as a p-median problem with tour
lengths as variable costs and solve it using an exact method. Routing is then solved using a savings
method. If the cost of the LRP with p depots is more than that with p–1, the procedure is stopped.
This can be viewed as a hierarchical method, since the routing costs are explicitly included in the
locational model. This approach of evaluating every possible move is only feasible if the number of
potential depots is fairly small, and hence may not be applicable in practice.
Srivastava and Benton (1990) and Srivastava (1993) present two very similar algorithms based
respectively on the moves “drop” and “add” in the locational phase. The routing phase in both is
solved using a savings algorithm. In the drop-based version, an “opportunity penalty” for not using
the best depot for each customer arc is used to select the depot for closure. (The add-based version
works symmetrically.) We note that this is a greedy procedure, as depots dropped/added cannot later
be reinstated/closed. Both versions compare favourably against a sequential method.
Chien (1993) uses route length estimation for the LRP. A number of heuristics are proposed,
consisting of combinations of the following modules. Initial solutions may be generated randomly,
with routing costs either calculated fully or estimated using one of two formulae designed and tested
11
by the author. Initial solutions may also be found using a modified closest-depot rule for the location
phase and a savings method for the routing phase. Solutions are then improved upon by a sequence of
operations (moving a group of customers from one route to another, inserting a customer from one
route to another, swapping two customers, reassigning all customers of a depot to another depot). A
comparison of the various combinations is given.
Nagy and Salhi (1996a) propose a hierarchical method called “nested method”. Nested methods
consist of a local search locational algorithm that refers to a routing method when evaluating
neighbouring solutions. The locational algorithm is based on tabu search and an add/drop/shift
neighbourhood. Thus, decisions made previously can be reversed, unlike in Srivastava (1993). After
each move, the routing solution is fully evaluated using a multi-depot VRP algorithm. However, the
costs of neighbouring solutions are approximated based on the observation that the impact of a change
in location is limited to a “region” and hence the routes are only recalculated for a limited area.
Various “region” shapes, based on computational geometrical ideas, are investigated. The method is
compared favourably against a sequential method. The above concept is taken further by Nagy and
Salhi (1996b), where costs of possible moves are approximated using route length estimation
formulae developed by the authors. The basic estimation formula can be enhanced by continued
comparison with actual routing costs and its parameters adjusted as necessary. Using such formulae
improves the speed of the algorithm considerably, leaving more time to finding better solutions.
Nested methods have subsequently been extended to the many-to-many location-routing problem
(Nagy and Salhi (1998), see 6.2.) and to the dynamic LRP (Salhi and Nagy (1999), see 5.3.). A
similar concept is also used by Tuzun and Burke (1999), who employ tabu search in both the location
and routing phases but evaluate neighbouring moves according to the sum of the depot-to-customer
distances, thus relying on a cruder guidance than Salhi and Nagy (1996a, 1996b). Their method
improves on the results of Srivastava (1993). However, no comparison with Salhi and Nagy (1996a,
1996b) is given.
Lin, Chow and Chen (2002) allow vehicles to take multiple trips. First, the minimum number of
facilities required is determined. Then, the VRP solution is completely evaluated for all combinations
of facilities. Vehicles are allocated to trips by completely evaluating all allocations. If the best
routing cost found is more than the setup cost for an additional depot, the algorithm moves on to
evaluating all sets of facilities that contain one more depot. The applicability of this method is limited
as it relies on evaluating what may well be a large number of depot configurations.
Albareda-Sambola, Díaz and Fernández (2005) apply an ingenious graph transformation to the
LRP. An initial solution is found via the linear programming relaxation of their model. The
locational neighbourhood is based on the moves add, drop and shift. However, no reference is made
to routing when evaluating possible moves. Infeasible solutions are allowed and a penalty term is
included in the objective function for depot capacity constraint violation. The overall algorithm is
encompassed in a tabu search framework, whereby the locational and routing routines correspond to
the diversification and intensification phases.
Melechovský, Prins and Wolfler Calvo (2005) consider an LRP with non-linear depot costs.
Initial solutions are found randomly or by a clustering-based procedure. The improvement phase
utilises a variable neighbourhood structure, based on moving a chain of customers from one route to
another. The widest neighbourhood in this structure, by reallocating all customers of a depot,
corresponds to a locational neighbourhood. (In this aspect it is somewhat similar to Chien (1993),
who also constructed the locational neighbourhood by reallocating all customers of a depot.) Within
each level of neighbourhood, tabu search is used to avoid local optima.
Hierarchical methods are also used for other types of LRPs. All the papers on the many-to-many
location-routing problem (see 6.2) rely on a hierarchical solution structure. They are also used in
solving stochastic and dynamic LRPs (see section 5) by Salhi and Nagy (1999), Liu and Lee (2003)
and Albareda-Sambola et al. (2007), while Soiud et al. (2007) (see 6.4) implemented a hierarchical
heuristic to solve the road-train routing problem.
12
Finally, we note that in principle, methodologies that either completely evaluate all possible depot
combinations or investigate a number of given depot configurations could be classified as hierarchical.
These include the pioneering works of Watson-Gandy and Dohrn (1973) and Or and Pierskalla
(1979), the dynamic LRP heuristic of Laporte and Dejax (1989) (see 5.3) and the many-to-many
problem of Lischak and Triesch (2007) (see 6.2).
5. Stochastic and dynamic problems
The only thing that is certain in life is that it is full of uncertainties! It is often unreasonable to
assume that all the parameters of an LRP problem are unchanging and known precisely. Although
taking uncertainty into account presents additional difficulties, there are a large number of papers of
the stochastic LRP. Most of these are devoted to the special case of one depot and one vehicle,
known as the travelling salesman location problem. We also note that all the papers in the literature
consider customer demand as the single input subject to stochastic variation.
Before reviewing these papers, we would like to comment on some striking characteristics of this
problem version. Compared to the deterministic LRP, a large proportion of papers uses an exact
method of solution. Another striking aspect is that for most heuristics proposed the authors furnish its
worst-case bound or optimality gap and its computational complexity. A summary of the most recent
major papers for each problem version is presented in Table 5.
Problem type
Travelling salesman location
Probabilistic TS location
Pickup delivery location
General stochastic LRP
Stochastic LRP with inventory
Dynamic LRP
Solution method
Exact: graph theoretical
Exact: graph theoretical
Graph theoretical heuristic
Hierarchical heuristic
Hierarchical heuristic
Clustering heuristic
Hierarchical heuristic
Paper
McDiarmid (1992)
Averbakh et al. (1994)
Mosheiov (1995)
Albareda-Sambola et al. (2007)
Liu and Lee (2003)
Chan et al. (2001)
Salhi and Nagy (1999)
Facilities
Not given
Not given
Customers
Not given
Not given
1
10
20
9
400
80
100
200
52
400
Table 5. Summary of recent papers on stochastic and dynamic problems.
5.1. Travelling salesman location problems
Travelling salesman location problems consider a set of customers, each time interval a subset of
which present a demand for service. This is not known in advance but is described by a probability
distribution. The objective is to locate the home base of the salesman such that the expected tour
length is minimised. Two models have been considered in the literature:
(1) For each subset of customers a tour is constructed.
(2) An a priori tour through all customers is found, and the actual tour will each day skip those
customers that do not require service.
We note that solving the first model requires more computational effort than the second one. It is also
possible that in practice one does not wish to reoptimise daily, in order to maintain regularity of
service. Nevertheless, the results of this model could serve as lower bounds to the second one. From
now on, we will use the expression “travelling salesman location problem” (TSLP) for the first model
only. The second model will be referred to as the probabilistic travelling salesman location problem
(PTSLP).
The travelling salesman location problem was introduced to the literature by Burness and White
(1976). They propose an iterative solution procedure where the facility is relocated in each step as the
median of the first and last customers in all possible TSP tours. Planar location with Euclidean or
rectilinear distances is assumed. Berman and Simchi-Levi (1986) prove the fundamental theorem of
13
TSLP namely that at least one of the optimal solutions is located on a vertex of the network. An
optimal polynomial-time algorithm for tree networks is also presented. Simchi-Levi and Berman
(1988) propose a polynomial-time heuristic for general networks, utilising an approximation formula
for the lengths of the TSP tours involved. This work is extended to planar location with Euclidean or
rectilinear distances by Simchi-Levi and Berman (1987), while its worst-case bound is improved by
Bertsimas (1989). McDiarmid (1992) considers the TSLP with fewer assumptions about probability
distributions and presents a linear-time algorithm for tree networks.
Mosheiov (1995) introduces the pickup delivery location problem, which combines the TSLP and
the TSP with pickups and deliveries. Furthermore, in this problem, apart from the customer set being
subject to variation, their demands are stochastic variables. The author extends the TSLP heuristics of
Berman and Simchi-Levi (1986) and Simchi-Levi and Berman (1988) to cater for this case and also
proposes a heuristic based on ranking customers.
The probabilistic travelling salesman location problem was introduced to the literature by Berman
and Simchi-Levi (1988). It is based on the underlying probabilistic travelling salesman problem,
which entails finding a tour that minimises the expected distance travelled by a salesman who follows
the tour but skips customers not requiring service. The authors present a branch-and-bound algorithm
for the PTSLP on a network. Bertsimas (1989) reduces the PTSLP to solving n probabilistic
travelling salesman problems and proposes two polynomial-time heuristics: a nearest neighbour based
one for networks and a spacefilling curve based one for the PTSLP on a Euclidean plane.
Only a few works investigate travelling salesman locations problems with non-standard or
multiple objectives. Averbakh and Berman (1995) introduce the multiobjective probabilistic salesdelivery location problem, a stochastic extension of their earlier paper (Averbakh and Berman (1994),
see section 3). Polynomial-time algorithms for tree location are presented. This is taken further by
Averbakh, Berman and Simchi-Levi (1994), who consider a variety of objectives for the probabilistic
travelling salesman location problem, such as minimising total or average waiting times or total tour
length, and show how their previous heuristics can be adapted to handle all these objectives.
A closely related problem is the travelling repairperson location problem, introduced by Jamil,
Batta and Malon (1994). This problem combines location with real-time routing. It involves a
travelling repairperson who sets out from his/her home base to service randomly arising calls,
travelling from customer to customer in the order he/she receives calls for service, only returning to
the home base once there are no customers awaiting service. The objective is to find a home base
location that minimises the average response time to a call for service. As routes are constructed by
emerging customer demand and cannot be planned in advance, this problem is quite rightly modelled
as a queueing-location rather than a location-routing problem. (See Berman and Krass (2001) and
Boffey et al. (2007) for reviews of this field.)
5.2. Stochastic location-routing with multiple vehicles
There are relatively few papers that consider more than one vehicle and they treat different
problem versions.
In the problem considered by Laporte, Louveaux and Mercure (1989), both depot locations and a
priori routes must be planned before the exact level of demand is known. This may result in routes
exceeding the vehicle capacity, known as a route failure. If this occurs, the vehicle returns to the
depot prematurely and then resumes service to the remaining customers. The cost of this additional
journey can be viewed as a penalty. The authors’ objective function is to minimise depot and a priori
route costs. This is subject to one of the following two constraints: (a) a limit on the probability of
route failure or (b) a limit on the expected penalty of a route. Solution is by relaxation of the
connectivity and integrality constraints, branch-and-bound and reintroduction of violated constraints.
Simchi-Levi (1991) extends the TSLP to several capacitated salesmen and extends the
fundamental theorem to this case, that is, at least one of the optimal solutions for the problem is on a
14
node of the network. A polynomial-time heuristic is presented for the case of networks and then
modified to cater for planar location with rectilinear or Euclidean distances.
Liu and Lee (2003) consider a stochastic customer demand and include inventory costs in the LRP.
An initial solution is found by clustering the customers, based on an increasing order of their marginal
inventory costs. For each cluster, the depot is located nearest to the centre and a TSP is solved. Then,
a hierarchical improvement method is used based on the moves drop and shift for the locational phase.
Both routing and inventory costs are fully evaluated for possible moves, thus the procedure is much
slower than “nested methods” that use route length estimation or other means of reducing the
computational burden.
In the problem studied by Albareda-Sambola et al. (2007), both depot locations and a priori routes
are designed before demand is known. The a posteriori routes may then omit some customers, if the
total demand is such that the vehicle capacity would be exceeded. Unserviced customers result in a
penalty. The objective function consists of the sum of depot costs, expected costs of a posteriori
routes and expected penalty costs. The authors also use approximations for the latter two costs. An
initial solution is found by a sequential location-allocation-routing heuristic. Then, this solution is
improved upon by local search, consisting of add and swap moves for location and insert, swap and
2-opt for routing.
For an excellent exposition and a more detailed review of stochastic location-routing, including
travelling salesman location problems, see Berman, Jaillet and Simchi-Levi (1995).
5.3. Dynamic location-routing
Dynamic problems divide the planning horizon into multiple periods. Normally within the
planning horizon there is some uncertainty about some of the parameters (typically the customer
demands), hence dynamic problems are related to the stochastic problems discussed above. We
consider dynamic location-routing a very important area of the LRP. This is because the static
(single-period) LRP is very much prone to the criticism that the planning horizons of the location and
routing subproblems do not match. By considering a planning horizon for facility location that
contain shorter planning intervals for route planning, dynamic LRPs are a much better model of
real-life location problems with routing aspects and provide an important means of refuting the above
criticism.
We may distinguish between two types of dynamic problems. In one, the depots are located
sequentially. In the other, the depots are located at the beginning of the planning horizon and vehicle
routes vary with the variations in customer demand. The former case is more applicable if demand is
increasing and the latter if demand is fluctuating.
In the problem studied by Nambiar et al. (1989) a rapid increase in supply was foreseen. Thus, the
authors’ solution provided a sequence of depot locations to be opened at different times. An
interesting consideration is that they allowed a factory to be closed down when another was opened
and to be re-opened later, in line with the real-life situation under investigation. However, the routing
considerations were chiefly neglected, thus their method cannot strictly be classified as an LRP.
Laporte and Dejax (1989) also consider multiple planning periods, whereby in each period both
the locations and the routes may be changed. They present an ingenuous network representation of
the problem, where some of the arcs are “spatial” vehicle routing arcs while some others are
“temporal” arcs representing the transformation from one time period to another. The resulting
network optimisation problem is solved to optimality following the procedure of Laporte et al. (1988).
The authors also present a heuristic method. For each time period and for each possible depot
configuration, distribution costs are estimated using a route length estimation formula. These costs
and the costs of transition from one configuration to another are represented on a network and a leastcost path problem is solved providing the solution as a sequence of depot configurations. The
applicability of this heuristic is limited by its exponential computational complexity.
15
Salhi and Nagy (1999) assume that the depots are fixed throughout the planning horizon but the
vehicle routes change following changes in customer demand. It is also assumed that the customer set
does not change. A number of solution approaches are investigated:
(a) a locational decision is made on the basis of average (forecasted) demands,
(b) an LRP is solved for each time period (as this violates the assumption of fixed depots, it serves as
a lower bound),
(c) a locational decision is made by selecting one of the set of solutions found in (b) and
(d) a locational decision is created using some of the depots featuring in the set of solutions found in (b).
The above scenarios were all evaluated using the method of Nagy and Salhi (1996b). It was found
that the solution found in (d) is very close to the lower bound and thus this selection rule provides a
good way of finding a set of depots that behave well under changing conditions.
Chan, Carter and Burnes (2001) use a clustering heuristic to investigate approaches (a) and (b),
classifying them as a priori and a posteriori approaches. They conclude that forecasted demands are
a reasonable approximation. Ambrosino and Scutellà (2005) consider a multi-level LRP (see 6.4) and
apply commercial software to their ILP formulation.
6. Problems with non-standard hierarchical structure
In the location-routing problems discussed thus far, we assumed a structure of facilities servicing a
number of customers, who are connected to their depot by means of vehicle tours. No routes connect
facilities to each other. (Some LRP papers consider a central pre-located facility to which all facilities
to be located must be connected: as this connection is by means of a direct link, its cost can be
included in the facility costs and no tour planning is involved. Thus, this is a negligible variation.)
This section will look at four problem versions, with very different hierarchical structures: some
simpler and some more complex than the LRP models discussed so far. Some of them may even be
considered as falling outside the definition of LRP, but are included for the sake of completeness.
They differ from each other and from the standard LRP according to whether tour planning is
involved in the various layers:
(1) the transportation-location problem does not involve tour planning,
(2) the many-to-many location-routing problem involves tour planning between facilities and
customers but also involves inter-facility routes,
(3) the vehicle routing-allocation problem involves tour planning between facilities but not between
facilities and customers and
(4) the multi-level location-routing problem involves tour planning at both layers and may even
consider more than one level of facility.
We note that all but one of the papers in this section are deterministic and static. Both exact and
heuristic approaches are discussed. The latest papers on each problem version are listed in Table 6.
Problem type
Many-to-many LRP
Solution method
Hierarchical heuristic
Vehicle routing–allocation
Exact: branch-and-cut
Sequential heuristic
Exact: branch-and-bound
Hierarchical heuristic
Multi-level LRP
Road-train routing
Paper
Nagy and Salhi (1998)
Wasner and Zäpfel (2004)
Labbé et al. (2005)
Murty and Djang (1999)
Ambrosino and Scutellà (2005)
Souid et al. (2007)
Facilities
249
10
150
331
28
30
Customers
249
2042
150
331
135
100
Table 6. Summary of recent papers on problems with non-standard hierarchical structures.
16
6.1. The transportation-location problem
The transportation-location problem (TLP) combines the problem of locating facilities with that
of transporting goods between supply and demand points. Each origin-to-destination route is a path
(not necessarily a simple path) through the facilities to be located. This problem is also known as the
transshipment-location problem or the path location-routing problem. It occurs especially frequently
in hazardous material transportation, where it is indeed more sensible not to have stops. In fact, all
the combined hazardous LRP models in the literature are, in fact, TLPs. While some authors classify
these as falling within the field of LRP, we do not do so, as these papers lack the essential
characteristic of tour planning. Thus, we present here merely a brief summary.
Hazardous TLPs have been addressed by Zografos and Samara (1989), ReVelle, Cohon and
Shobrys (1991), List and Mirchandani (1991), Stowers and Palekar (1993), Boffey and Karkazis
(1993), Kulcar (1996), Giannikos (1998) and Cappanera, Gallo and Maffioli (2004). These papers
adopt a multiobjective approach, trying to find an appropriate balance between minimising
transportation costs and minimising risk. For a review of hazardous material location and routing, see
List et al. (1991), Boffey and Karkazis (1995) and Erkut and Verter (1995).
There are relatively few papers on transportation-location problems other than hazardous
applications. We refer the readers to the pioneering papers of Maranzana (1964) and Cooper (1972),
the multiobjective approach of Ogryczak, Studziński and Zorychta (1989, 1992), the multiproduct
problem of Hindi and Basta (1994) and the hub location type problem of Aykin (1995).
6.2. The many-to-many location-routing problem
Nagy and Salhi (1998) introduce the many-to-many location-routing problem (MMLRP). In this
problem, several customers wish to send goods to others. In the most general case, it is assumed that
each customer sends a different commodity to every other customer. This corresponds to the case of
postal flow between localities. A network of hubs is to be located, taking into account routing costs.
(Inter-hub routes are assumed to be direct while hub-to-customer routes are multi-stop.) Just as the
LRP is an approach to locate facilities, the MMLRP is an approach to locate hubs. It is noted that as
customers can both send and receive goods, a pickup-and-delivery routing method is required to find
routing costs. This is harder to solve than the VRP, as a fluctuating load on the vehicles makes
feasibility checks harder to perform. A hierarchical heuristic solution framework, based on the
concept of “nested methods” is presented. We note that a number of logistics problems are special
cases of the MMLRP, such as the hub location problem (if full-truckload routes are assumed), the
freight transport problem (if hubs are fixed) and of course the LRP (if there is no inter-hub flow and
either all deliveries or all pickups are zero). This type of problem has application in the design of letter
or parcel delivery systems and in fact all the other papers on this problem concern parcel delivery
applications. Since they tackle practical applications, no computational comparisons to other papers
are given. All authors chose hierarchical solution methods.
Bruns, Klose and Stähly (2000) study a problem arising in the parcel delivery operations of a
postal service. In this system, parcels travel directly from post offices to parcel processing centres and
thence to delivery bases. From there, they are delivered to customers by vehicles making multiple
stops. (Processing centres themselves can act as delivery bases.) The problem at hand was to
determine the locations of the delivery bases, their allocation to processing centres and the allocation
of customer areas to delivery bases. (The processing centre locations are predetermined.) As in this
problem the flow from customers (post offices) to delivery bases is via paths and is separate from the
flow from bases to customers, the authors are able to reduce their model to an LRP. They investigate
the costs incurred and find appropriate approximations for them; in the case of routing costs, this
involves designing a route length estimation formula. Thus, the authors are able to further reduce the
LRP to a simple plant location problem, where routing costs are subsumed into the assignment costs.
This problem is then solved using branch-and-bound.
17
The problem studied by Wasner and Zäpfel (2004), posed by a parcel delivery service provider, is
more closely related to the MMLRP. Vehicles perform both deliveries and pickups and both their
home base location and their routes need to be determined. However, all inter-hub flow must travel
through a central hub, rather than allowing all hubs to be directly connected to each other. (As this
central hub has already been located, this is a minor variation.) A hierarchical solution approach is
adopted. Depot locations are found by an add/drop/shift heuristic. Vehicle routes are also found
heuristically. The overall method is driven by four feedback mechanisms: (1) varies the allocation of
customers to depots, (2) combines “neighbouring” vehicle tours, (3) changes the location of the depots
(shift) and (4) changes the number of depots (add/drop).
Lischak and Triesch (2007) also investigate a parcel delivery MMLRP. In their problem, parcels
between depots may travel directly or via one or two hubs. The majority of the depot locations have
been predetermined by the client. Apart from the location of the remaining depots, also the “class”
(throughput capacity) of all the depots needs to be determined. The authors choose to completely
enumerate all possible depot combinations, using a pickup-and-delivery heuristic to compute routing
costs. An attempt to incorporate route length estimation is also made.
Finally, we note that the problems investigated by Min (1996) (see 4.2.) and Gunnarsson et al.
(2006) (see 6.3.) also involve several sources or destinations, but do not involve tour planning in
either the delivery or the collection routes.
6.3. Vehicle routing-allocation problems
In the above papers, inter-hub routes were direct while hub-to-customer routes involved tours. We
could turn this around and consider tours involving hubs and customers being allocated to hubs
directly. It is debatable whether such problems should be considered part of the LRP, since no tour
planning is involved at customer level, but they are both interesting and sufficiently closely related to
be included here. Beasley and Nascimento (1996) name this problem the vehicle routing-allocation
problem (VRAP) and give a review of related works. However, different authors use different names
and create slightly different problem versions.
Nambiar et al. (1981) consider the allocation of rubber smallholders to rubber collection stations
together with designing the routes of collection vehicles, but their solution algorithm ignores the
allocation issue. Labbé and Laporte (1986) solve a problem of locating postboxes. They seek to
minimise a linear combination of vehicle routing costs (that of the postal collection van) and
customers’ inconvenience costs (these depend on the sum of distances between customers and their
allocated postbox). A sequential and an iterative heuristic are proposed. Murty and Djang (1999)
investigate a complex military logistics problem centred on a VRAP. Mobile training simulators
traverse training sites, while army units travel directly from their home bases to training sites. A
sequential solution method is presented, based on set covering for locating the training sites and a
constructive heuristic for routing the simulators. Labbé et al. (2005) use a branch-and-cut method to
solve a VRAP (called the median cycle problem) where the objective is to minimise routing cost
subject to an upper bound on the allocation cost. An interesting practical problem, similar to the
VRAP, is tackled by Gunnarsson, Rönnqvist and Carlsson (2006). Routes from sources to hubs may
involve stops (at other sources or hubs) but routes from hubs to customers are always direct. A
heterogeneous fleet consisting of smaller and larger ships, trains and lorries is used. The authors
simplify the problem by considering only a small subset of all possible multi-stop routes and
associating a variable with each route – thus, their solution algorithm involves no tour planning. The
relaxed problem is solved using a commercial ILP solver. Furthermore, three heuristics, based on
removing and later reintroducing some of the variables and constraints, are proposed.
The above papers provide a bridge from LRP to the field of extensive facility location, since
instead of viewing the problem as that of designing a route we could view it as that of locating a cycle.
Extensive facility location concerns locating a dimensional structure, for example a path or a cycle on
18
a network. This falls outside the scope of this paper – we refer to Mesa and Boffey (1996), Labbé et
al. (1988) and Díaz-Báñez et al. (2004) for reviews.
6.4. Multi-level location-routing problems
An imaginative reader may by now have imagined the situation where routing occurs both at hub
and customer level. Others may well think that this would be a theoretical construct unlikely to occur
in practice – however, this is not the case. In the two-level location-routing problem, introduced by
Jacobsen and Madsen (1980) and Madsen (1983), newspapers are delivered from the factory to
transfer points and from these to the customers. The problem consists of:
(a) determining the locations of transfer points,
(b) designing a vehicle route through these points (known as a primary tour),
(c) allocate the customers to transfer points (or directly to the factory),
(d) designing vehicle routes for each of the above customer clusters (known as secondary tours).
We note that the two-level LRP can be viewed as an LRP extension, since it combines the problem of
finding the transfer point locations with the customer delivery routing problem. It can also be viewed
as an extension to the VRAP where allocation is achieved via tours rather than by direct links. An
unusual aspect of the problem is that transfer points can be relocated with little expense. The
“tree-tour” heuristic proposed by the authors is based on the observation that if one deletes the last arc
from each route, the problem becomes similar to a Steiner tree problem. This tree is constructed by a
greedy one-arc-at-a-time procedure. The authors also put forward two sequential heuristics.
The road-train routing problem, introduced by Semet and Taillard (1993), can also be viewed as a
two-level LRP. This problem concerns designing a route for a vehicle, called a road-train, that is
composed of two parts, a truck and a trailer. Some of the customers are not accessible to this vehicle
and thus the trailer is detached and left at a customer location (called a “root”) while the truck visits a
subset of these customers, returning to pick up the trailer. The route of the road-train corresponds to
the primary tour and the routes run by the truck alone to secondary tours. The difference from the
Jacobsen and Madsen (1980) problem is that here some customers can be served directly by the
primary tour. However, similarly to it, facility costs are negligible and the trailers can be parked at
different locations each day. Hence, this problem lacks the usual strategic aspect of LRPs. We note
that different authors tackle slightly different versions of the problem.
Semet and Taillard (1993) find an initial solution using a sequential procedure and improve this by
a tabu search method, where customers (truck or trailer) are reallocated. In terms of the LRP, this
method does not distinguish between locational and routing moves. All trailer customers are forced to
be located on the primary tour, which may not lead to the best solution. Semet (1995) proposed a
clustering-based solution method. First, customers are allocated to roots: this is formulated as an
assignment problem and solved using Lagrangean relaxation. Then, the resulting travelling salesman
problems are solved. Gerdessen (1996) assumes that all customers have unit demand and each trailer
is parked exactly once. Initial solutions are found using a number of sequential heuristics. These are
then improved by a selection of VRP improvement heuristics. Chao (2002) uses a cluster-first routing
second initial solution and a tabu search improvement phases with customer reallocation moves. Both
papers – similarly to Semet and Taillard (1993) – do not distinguish between locational and routing
moves. Souid, Hanafi and Semet (2007) allow more than one subtour (truck route) to originate from a
root. Similarly to Melechovský et al. (2005), a hierarchical variable neighbourhood structure is used,
with the smaller neighbourhoods relating to relocating customers within or between secondary tours
and the widest one is defined by the moves of adding or dropping a root.
Perhaps the most complex location-routing problem is that of Ambrosino and Scutellà (2005).
They consider a four-level LRP, where level 1 is the plant, level 2 consists of distribution centres,
level 3 contains transfer points and some (mainly large) customers and level 4 consists of customers.
Tour planning is present from level 2 downward. The authors also introduce inventory considerations.
Both static and dynamic problem cases are treated. Furthermore, a heterogeneous fleet is allowed.
The problem is formulated as an ILP and solved to optimality using commercial software.
19
7. Suggestions for future research
It may be interesting to begin by looking at what the authors of the past three LRP reviews
considered promising future research directions and investigating how much of their suggestions have
been realised.
The main recommendations of Balakrishnan, Ward and Wong (1987) can be summarised as:
(a) Investigating multiple planning periods. This issue, also referred to as the dynamic LRP, has been
considered explicitly in Nambiar et al. (1989), Laporte and Dejax (1989), Salhi and Nagy (1999)
and Ambrosino and Scutellà (2005). The Salhi and Nagy (1999) study also provided an important
validation of LRP as a research field, answering criticisms arising from the issue of different
planning horizons. However, a more in-depth investigation of dynamic multi-period problems
would enhance the standing of location-routing as a research area.
(b) Explicitly comparing sequential and combined methods. This has been achieved by Salhi and
Rand (1989). It was found that the use of a sequential method leads to worse solutions than a
combined method. These results provide an important validation to the field of location-routing.
(c) Using analytical formulae in a hierarchical framework. Such “route length estimation” formulae
were derived and then used to solve the LRP by Laporte and Dejax (1989), Chien (1993), Nagy
and Salhi (1996b), Bruns, Klose and Stähly (2000) and Lischak and Triesch (2007).
Laporte (1989) identifies the following four “promising research areas”:
(d) Development and systematic analysis of LRP heuristics. A significant number of articles appeared
proposing various LRP heuristics. However, a proper analysis and comparison of these is sadly
lacking from the literature, with the notable exception of travelling salesman location problems.
(e) Development of new exact methods based on Lagrangean relaxation or dynamic programming.
No such paper has appeared, although two LRP subproblems were solved by Lagrangean
relaxation by Semet (1995) and Bruns, Klose and Stähly (2000). In our view Lagrangean
relaxation is likely to be a promising and challenging research approach.
(f) A study of hierarchical LRPs. Such studies were investigated in a theoretical framework (called
“nested methods”) by Nagy and Salhi (1996a, 1996b, 1998). Practical multi-level LRP versions
were also tackled by Bruns, Klose and Stähly (2000), Wasner and Zäpfel (2004) and Lischak and
Triesch (2007).
(g) A study of dynamic LRPs. [Equivalent to suggestion (a) of Balakrishnan et al. (1987).]
Min, Jayaraman and Srivastava (1998) list eight “future research directions”:
(h) Stochasticity. This issue has since been considered by Chan, Carter and Burnes (2001), Liu and
Lee (2003) and Albareda-Sambola et al. (2007).
(i) Time windows. Only Semet and Taillard (2003) considered this issue, for the special case of the
road-train routing problem. Perhaps as time windows relate to a much smaller time horizon than
facility location, this “horizon mismatch” may have deterred researchers from including this
aspect. (Time windows are usually subject to more fluctuation than the routes themselves.)
(j) Multiple periods. [Equivalent to suggestion (a) of Balakrishnan et al. (1987).]
(k) Multiple objectives. To date, the only Hamiltonian LRP with multiple objectives considered in the
literature is the sales-delivery location problem of Averbakh and Berman (1994, 1995). We agree
that this is an important issue to consider.
(l) Vertical integration. This means consideration of both delivery and pickup traffic and has now
been considered by Nagy and Salhi (1998), Wasner and Zäpfel (2004) and Lischak and Triesch
(2007).
(m)Horizontal integration. This essentially means the integration of inventory aspects into the LRP.
This has now been considered by Liu and Lee (2003) and Ambrosino and Scutellà (2005).
(n) Benchmarks for solution efficiency. Sadly, the LRP literature is very fragmented and no widely
accepted benchmarks exist. (There are collections of instances available from Barreto (2003) or
Klose (2005).) In the entire subject literature, only four papers (namely Wu et al. (2002), Hansen
et al. (1994), Nagy and Salhi (1996b) and Tuzun and Burke (1999)) give computational
comparisons with their peers, making it nearly impossible to properly judge solution quality.
20
(o) Application to real-world problems. A number of such studies have since appeared, confirming
the applicability of location-routing (see subsection 1.3). We hope that practitioners realise the
importance of using LRP models and methods to solve their location problems! We believe that
the applicability of LRP would increase if LRP methods were coupled with an easy-to-use
interface (perhaps in a GIS framework).
The above observations paint a mixed picture. Although significant progress has been made,
certain challenges have not yet been taken up by the locational analysis community.
In the following, we give some suggestions for future research, which we believe to be worthwhile
considering. (Some of these were also suggested by previous reviews, as indicated by the references
to the above list given in brackets.) The authors are currently working on some of these problems.
1. Hierarchical methods and use of route length formulae. [(c), (f)] The authors’ view is that routing
plays a subservient role in LRP and approximation formulae can often be used instead of vehicle
routing algorithms within the search to speed up the process. One possible approach is to refer to
the routing stage explicitly whenever necessary and use approximation in other steps. Thus, we
feel that continued testing and further development of such formulae would be desirable. In
particular, formulae should be extended to cater for vehicle routing extensions such as fleet mix or
pickup and delivery. Another interesting topic would be the development of an estimation formula
for the arc routing problem. (This would be very different from route length estimation in node
routing, since it is the deadheading distance that needs to be estimated.)
2. Dynamic and stochastic problems. [(a), (g), (j)] The most important criticism of location-routing is
that during the planning horizon of facilities, the customer demands may fluctuate and even the
customer set may change. Thus, this type of problems merits further investigation. In particular,
little research has been done on the dynamic LRP where depots are located in a sequence. We
think that this problem has real-world applicability – we can envisage a growing demand or
coverage but the pattern of this growth is not known with certainty. (In fact, one of the papers
proposing this problem was a case study.) This problem could best be solved using robustness
analysis. This methodology is geared towards sequential decision-making and strives to preserve
flexibility in an uncertain environment. Thus, it would be very suitable for this type of dynamic
LRP. Another interesting problem would be to investigate problems where arc lengths, rather than
customer demands, are stochastic. This would model real-life situations better, as travel times tend
to depend on travel conditions.
3. Planar location-routing. Looking back at our review, we can observe that all papers, with the
exception of some TSLP and round-trip location papers and the recent Schwardt and Dethloff
(2005) and Salhi and Nagy (2007) papers, deal with discrete location. Yet, in the general location
literature, there is a considerable proportion of articles on continuous location. It is puzzling why
there are hardly any works on planar LRP. Clearly, the LRP is more applicable to discrete
problems as the vehicle routing aspect assumes an underlying road network. However, a planar
problem could be construed where the customers are on a road network but the facility would be
located on a greenfield (brownfield) site.
4. Integrated methods in logistics. [(l), (m)] The authors are great believers in integrated methods in
logistics. Combining location-routing with other aspects of logistics would be an interesting
avenue of research. We are particular interested in solving the location-routing-inventory and the
location-routing-packing problems.
5. Multiobjective LRP. [(k)] Very little has been done in this area, yet it can clearly be applied to a
variety of problems. On one hand, the problem of hazardous material transportation attracted
considerable attention in the literature. Some of these works address issues of location and routing
in a combined framework, however the underlying routes are not of a multi-stop nature. It would
be an interesting problem to solve combined location-routing problems with multi-stop routes
where the material transported is either hazardous or obnoxious. On the other hand, we would like
21
to see development of further models for delivery-man or similar problems. For example, the
newspaper delivery problem of Jacobsen and Madsen (1980) could be solved using a
multiobjective approach, minimising total tour length on one hand and the sum of arrival times at
customers on the other.
6. Competitive LRP. Competitive location has a considerable literature, giving rise to interesting
connections with game theory. Yet, there are no papers in competitive location-routing. One
possible explanation for this is that as competitive location focuses on user choice, it is more
reasonable to assume that users travel directly to facilities, and hence no routing considerations are
required. Perhaps a competitive VRAP, such as the following, is easier to imagine than a
competitive LRP. Two transport companies wish to locate circular routes. Customers may just
choose the stop nearest to them, or for each possible origin-destination pair they may choose the
company that minimises the sum of origin-to-stop and stop-to-destination distances.
7. Heuristics for Eulerian location. This has obvious applications to locating postal delivery, road
maintenance or waste collection depots. There are only two papers on this problem: one ignores
the interdependency of location and routing and the other ignores important constraints. It would
be interesting to see iterative and integrated heuristics developed for this problem.
8. Hybrid methodologies. Although the location-routing research community is small, research is
somewhat fragmented and a number of strands exist. We hope that this review may help in
identifying different methodologies and prompt researchers to see if they can be united in some
way. For example, the lack of feedback in clustering-based methods could be eliminated by
including them in an iterative framework. Or, such methods could provide better starting points for
a hierarchical local search algorithm. Another promising avenue would be the combination of
exact and heuristic methods, such as heuristic concentration.
9. Modelling complex situations. We also believe LRP and similar models should be applied to
complex situations. For example, it would be interesting to see an iterative or hierarchical solution
to the problem faced by Murty and Djang (1999) rather than the sequential approach adopted by
the authors. Similarly, the four-level static and dynamic LRPs introduced by Ambrosino and
Scutellà (2005) could be solved heuristically. The MMLRP is also a growing research field, with
already some papers published on postal and parcel delivery applications, however it could also be
applied to rail-hub location in combined road-rail transport systems. Location-routing type models
could be applied to the design of public transport networks, where both metro and connecting bus
lines need to be located together.
We may summarise our proposed research agenda as trying to solve all locational problems, where
appropriate, using the location-routing approach. This review can happily report on a much larger
body of work than its precursors, but location-routing is still a developing area. We hope that the
challenges posed here will arouse the interest of some of the readers and entice them to work in this
challenging and exciting research area!
Acknowledgements
We would like to express our thanks to Mari Albareda-Sambola, Damla Ahıpaşaoğlu and Sérgio
Barreto for sharing with us their bibliographical research. We are also very grateful to Paola
Cappanera and three anonymous referees for their constructive comments, which improved both the
content as well as the presentation of the paper.
22
References
Albareda-Sambola, M., 2003, Models and Algorithms for Location-Routing and Related Problems. PhD
dissertation, Catalonia Polytechnic University.
Albareda-Sambola, M., Díaz, J.A., Fernández, E., 2005, A compact model and tight bounds for a combined locationrouting problem. Computers and Operations Research 32, 407-428.
Albareda-Sambola, M., Fernández, E., Laporte, G., 2007, Heuristic and lower bound for a stochastic location-routing
problem. European Journal of Operational Research (to appear).
Ambrosino, D., Scutellà, M.G., 2005, Distribution network design: new problems and related models. European
Journal of Operational Research 165, 610-624.
Averbakh, I., Berman, O., 1994, Routing and location-routing p-delivery men problems on a path. Transportation
Science 28, 184-197.
Averbakh, I., Berman, O., 1995, Probabilistic sales-delivery man and sales-delivery facility location problems on a
tree. Transportation Science 29, 184-197.
Averbakh, I., Berman, O., 1999, A simple heuristic for m-machine flow-shop and its applications in routingscheduling problems. Operations Research 47, 165-170.
Averbakh, I., Berman, O., 2002, Minmax p-traveling salesmen location problems on a tree. Annals of Operations
Research 110, 55-62.
Averbakh, I., Berman, O., Simchi-Levi, D., 1994, Probabilistic a priori routing-location problems. Naval Research
Logistics 41, 973-989.
Aykin, T., 1995, The hub location and routing problem. European Journal of Operational Research 83, 200-219.
Baita, F., Ukovich, W., Pesenti, R., Favaretto, D., 1998, Dynamic routing-and-inventory problem: a review.
Transportation Research A 32, 585-598.
Balakrishnan, A., Ward, J.E., Wong, R.T., 1987, Integrated facility location and vehicle routing models: recent work
and future prospects. American Journal of Mathematical and Management Sciences 7, 35-61.
Barreto, S.S., 2004, Análise e Modelização de Problemas de Localização-Distribuição. PhD Dissertation, Aveiro
University.
Barreto, S.S., 2003, http://sweet.ua.pt/~iscf143/
Barreto, S., Ferreira, C., Paixão, J., Sousa Santos, B., 2007, Using clustering analysis in a capacitated locationrouting problem. European Journal of Operational Research (to appear).
Beasley, J., Nascimento, E.M., 1996, The vehicle routing-allocation problem: a unifying framework. TOP 4, 65-86.
Bednar, L., Strohmeier, E., 1979, Lagerstandortoptimierung und Fuhrparkeinsatzplanung in der KonzumgüterIndustrie. Zeitschrift für Operations Research 23, B89-B104.
Berger, R., 1997, Location-Routing Models for Distribution System Design, PhD dissertation, Northwestern
University.
Berman, O., Krass, D., 2001, Facility location problems with stochastic demands and congestion. In: Drezner, Z.,
Hamacher, H.W., Facility Location: Applications and Theory, Springer, Berlin, pp.329-371.
Berman, O., Jaillet, P., Simchi-Levi, D., 1995, Location-routing problems with uncertainty. In: Drezner, Z. (Ed.),
Facility Location: A Survey of Applications and Methods, Springer, New York, pp.427-452.
Berman, O., Simchi-Levi, D., 1986, Minisum location of a traveling salesman. Networks 16, 239-254.
Berman, O., Simchi-Levi, D., 1988, Finding the optimal a priori tour and location of a traveling salesman with
nonhomogenous customers. Transportation Science 22, 148-154.
Bertsimas, D.J., 1989, Traveling salesman facility location problems. Transportation Science 23, 184-191.
Billionnet, A., Elloumi, S., Grouz-Djerbi, L., 2005, Designing radio-mobile access networks based on synchronous
digital hierarchy rings. Computers and Operations Research 32, 379-394.
Boffey, B., Karkazis, J., 1993, Models and methods for location and routing decision relating to hazardous materials.
Studies in Locational Analysis 5, 149-166.
Boffey, B., Karkazis, J., 1995, Location, routing and the environment. In: Drezner, Z. (Ed.), Facility Location: A
Survey of Applications and Methods, Springer, New York, pp.453-466.
Boffey, B., Galvão, R., Espejo, L., 2007, Capacity effects in the location of facilities with immobile servers.
European Journal of Operational Research (to appear).
Bookbinder, J.H., Reece, K.E., 1988, Vehicle routing considerations in distribution system design. European Journal
of Operational Research 37, 204-213.
23
Branco, I.M., Coelho, J.D., 1990, The Hamiltonian p-median problem. European Journal of Operational Research
47, 86-95.
Bruns, A.D., 1998, Zweistufige Standortplanung unter Berücksichtigung von Tourenplanungsaspekten – Primale
Heuristiken und Lokale Suchverfahren, PhD Dissertation, Sankt Gallen University
Bruns, A., Klose, A., Stähly, P., 2000, Restructuring of Swiss parcel delivery services. OR Spektrum 22, 285-302.
Burness, R.C., White, J.A., 1976, The traveling salesman location problem. Transportation Science 10, 348-360.
Cappanera, P., Gallo, G., Maffioli, F., 2004, Discrete facility location and routing of obnoxious activities. Discrete
Applied Mathematics 133, 3-28.
Çetiner, S., 2003, An Iterative Hub Location and Routing Problem for Postal Delivery Systems. MSc dissertation,
Middle East Technical University.
Chan, A.W., Francis, R.L., 1976, A round-trip location problem on a tree graph. Transportation Science 10, 35-51.
Chan, A.W., Hearn, D.W., 1977, A rectilinear distance round-trip location problem. Transportation Science 11, 107123.
Chan, Y., Carter, W.B., Burnes, M.D., 2001, A multiple-depot, multiple-vehicle, location-routing problem with
stochastically processed demands. Computers and Operations Research 28, 803-826.
Chao, I-M., 2002, A tabu search method for the truck and trailer routing problem. Computers and Operations
Research 29, 33-51.
Chien, T.W., 1993, Heuristic procedures for practical-sized uncapacitated location-capacitated routing problems.
Decision Sciences 24, 995-1021.
Christofides, N., Eilon, S., 1969, Expected distances in distribution problems. Operational Research Quarterly 20,
437-443.
Christofides, N., Mingozzi, A., Toth, P., 1979, The vehicle routing problem. In: Christofides, N., Mingozzi, A.,
Toth, P., Sandi, C. (Eds.), Combinatorial Optimization, Wiley, Chichester, pp. 315-338.
Cooper, L., 1972, The transportation-location problem. Operations Research 20, 94-108.
Daganzo, C.F., 2005, Logistics Systems Analysis, Springer, Heidelberg.
Daskin, M.S., Coullard, C.R., Shen, Z-J.M., 2002, An inventory-location problem: formulation, solution algorithm
and computational results. Annals of Operations Research 110, 83-106.
Díaz-Báñez, J.M., Mesa, J.A., Schöbel, A., 2004, Continuous location of dimensional structures. European Journal
of Operational Research 152, 22-44.
Drezner, Z., 1982, Fast algorithms for the round trip location problem. IEE Transactions 14, 243-248.
Drezner, Z., Scott, C., Song, J-S., 2003, The central warehouse location problem revisited, IMA Journal of
Management Mathematics 14, 321-336.
Drezner, Z., Wesolowsky, G.O., 1982, A trajectory approach for the round-trip location problem. Transportation
Science 16, 56-66.
Erkut, E., Verter, V., 1995, Hazardous material logistics. In: Drezner, Z. (Ed.), Facility Location: A Survey of
Applications and Methods, Springer, New York, pp.467-506.
EWGLA, 2003, http://www.vub.ac.be/EWGLA/literature.html
Ferrer, L., Nagy, G., Wassan, N.A., 2007, The vehicle routing problem with packing considerations. (in progress)
Gerdessen, J.C., 1996, Vehicle routing problem with trailers. European Journal of Operational Research 93, 135-147.
Ghiani, G., Laporte, G., 1999, Eulerian location problems. Networks 34, 291-302.
Giannikos, I., 1998, A multiobjective programming model for locating treatment sites and routing hazardous wastes.
European Journal of Operational Research 104, 333-342.
Gunnarsson, H., Rönnqvist, M., Carlsson, D., 2006, A combined terminal location and ship routing problem.
Journal of the Operational Research Society (to appear).
Hansen, P.H., Hegedahl, B., Hjortkjær, S., Obel, B., 1994, A heuristic solution to the warehouse location-routing
problem. European Journal of Operational Research 76, 111-127.
Hennes, H., Hamacher, H.W., 2006, Integrated scheduling and location models: single machine makespan problems.
Studies in Locational Analysis (to appear).
Hindi, K.S., Basta, T., 1994, Computationally efficient solution of a multiproduct, two-stage distribution-location
problem. Journal of the Operational Research Society 45, 1316-1323.
Jacobsen, S.K., Madsen, O.B.G., 1980, A comparative study of heuristics for a two-level routing-location problem.
European Journal of Operational Research 5, 378-387.
24
Jamil, M., Batta, R., Malon, D.M., 1994, The traveling repairperson home base location problem. Transportation
Science 28, 150-161.
Klose, A., 2005, http://www.wiwi.uni-wuppertal.de/TSCFLRP.1321.0.html
Kolen, A., 1985, The round-trip p-center and covering problem on a tree. Transportation Science 19, 222-234.
Kulcar, T., 1996, Optimizing solid waste collection in Brussels. European Journal of Operational Research 90, 26-44.
Labbé, M., Laporte, G., 1986, Maximizing user convenience and postal service efficiency in post box location.
Belgian Journal of Operational Research, Statistics and Computer Science 26, 21-35.
Labbé, M., Laporte, G., Rodríguez-Martín, I., 1988, Path, tree and cycle location. In: Crainic, T.G., Laporte, G.,
(Eds.), Fleet Management and Logistics, Kluwer, Boston, pp.187-204.
Labbé, M., Laporte, G., Rodríguez-Martín, I., Salazar-González, J.J., 2005, Locating median cycles in networks.
European Journal of Operational Research 160, 457-470.
Labbé, M., Rodríguez-Martín, I., Salazar-Gonzalez, J.J., 2004, A branch-and-cut algorithm for the plant-cycle
location problem. Journal of the Operational Research Society 55, 513-520.
Laporte, G., 1988, Location-routing problems. In: Golden, B.L., Assad, A.A. (Eds.), Vehicle Routing: Methods and
Studies, North-Holland, Amsterdam, pp.163-198.
Laporte, G., 1989, A survey of algorithms for location-routing problems. Investigación Operativa 1, 93-123.
Laporte, G., Dejax, P.J., 1989, Dynamic location-routeing problems. Journal of the Operational Research Society 40,
471-482.
Laporte, G., Gendreau, M., Potvin, J-Y., Semet, F., 2000, Classical and modern heuristics for the vehicle routing
problem. International Transactions in Operational Research 7, 285-300.
Laporte, G., Louveaux, F., Mercure, H., 1989, Models and exact solutions for a class of stochastic location-routing
problems. European Journal of Operational Research 39, 71-78.
Laporte, G., Nobert, Y., 1981, An exact algorithm for minimizing routing and operating costs in depot location.
European Journal of Operational Research 6, 224-226.
Laporte, G., Nobert, Y., Arpin, D., 1986, An exact algorithm for solving a capacitated location-routing problem.
Annals of Operations Research 6, 293-310.
Laporte, G., Nobert, Y., Pelletier, P., 1983, Hamiltonian location problems. European Journal of Operational
Research 12, 82-89.
Laporte, G., Nobert, Y., Taillefer, S., 1988, Solving a family of multi-depot vehicle routing and location-routing
problems. Transportation Science 22, 161-172.
Lee, Y., Kim, S-i., Lee, S., Kang, K., 2003, A location-routing problem in designing optical internet access with
WDM systems. Photonic Network Communications 6, 151-160.
Levy, L., Bodin, L., 1989, The arc oriented location routing problem. INFOR 27, 74-94.
Lin, C.K.Y., Chow, C.K., Chen, A., 2002, A location-routing-loading problem for bill delivery services. Computers
and Industrial Engineering 43, 5-25.
Lischak, C., 2001, Standortplanung für einem Privaten Paketdienstleister, PhD dissertation, Technical University of
Rhineland-Westphalia Aachen.
Lischak, C., Triesch, E., 2007, Location planning for a parcel delivery service. (in review)
List, G.F., Mirchandani, P.B., 1991, An integrated network/planar multiobjective model for routing and siting for
hazardous materials and wastes. Transportation Science 25, 146-156.
List, G.F., Mirchandani, P.B., Turnquist, M.A., Zografos, K.G., 1991, Modeling and analysis for hazardous
materials transportation: risk analysis, routing/scheduling and facility location. Transportation Science 25, 100-114.
Liu, S.C., Lee, S.B., 2003, A two-phase heuristic method for the multi-depot location routing problem taking
inventory control decisions into considerations. International Journal of Advanced Manufacturing Technology 22,
941-950.
Madsen, O.B.G., 1983, Methods for solving combined two level location-routing problems of realistic dimensions.
European Journal of Operational Research 12, 295-301.
Maranzana, F.E., 1964, On the location of supply points to minimise transport costs. Operational Research Quarterly
15, 261-270.
Maze, T.H., Khasnabis, S., 1985, Bus garage location planning with dynamic vehicle assignments: a methodology.
Transportation Research B 19, 1-13.
McDiarmid, C., 1992, Probability modelling optimal location of a travelling salesman. Journal of the Operational
Research Society 43, 533-538.
25
Melechovský, J., Prins, C., Wolfler Calvo, R., 2005, A metaheuristic to solve a location-routing problem with nonlinear costs. Journal of Heuristics 11, 375-391.
Melkote, S., 1996, Integrated Models of Facility Location and Network Design, PhD dissertation, Northwestern
University.
Melkote, S., Daskin, M.S., 2001, An integrated model of facility location and transportation network design.
Transportation Research A 35, 515-538.
Mesa, J.A., Boffey, T.B., 1996, A review of extensive facility location in networks. European Journal of Operational
Research 95, 592-603.
Metters, R.D., 1996, Interdependent transportation and production activity at the United States postal service.
Journal of the Operational Research Society 47, 27-37.
Min, H., 1996, Consolidation terminal location-allocation and consolidated routing problems. Journal of Business
Logistics 17, 235-263.
Min, H., Jayaraman, V., Srivastava, R., 1998, Combined location-routing problems: a synthesis and future research
directions. European Journal of Operational Research 108, 1-15.
Moin, N.H., Salhi, S., 2007, Inventory routing problems: a logistical overview. Journal of the Operational Research
Society (to appear).
Mosheiov, G., 1995, The pickup delivery location problem on networks. Networks 26, 243-251.
Murty, K.G., Djang, P.A., 1999, The U.S. army national guard’s mobile training simulators location and routing
problem. Operations Research 47, 175-182.
Nagy, G., 1996, Heuristic Methods for the Many-to-Many Location-Routing Problem, PhD dissertation, University
of Birmingham.
Nagy, G., Salhi, S., 1996a, Nested heuristic methods for the location-routeing problem. Journal of the Operational
Research Society 47, 1166-1174.
Nagy, G., Salhi, S., 1996b, A nested location-routing heuristic using route length estimation. Studies in Locational
Analysis 10, 109-127.
Nagy, G., Salhi, S., 1998, The many-to-many location-routing problem. TOP 6, 261-275.
Nambiar, J.M., Gelders, L.F., Van Wassenhove, L.N., 1981, A large scale location-allocation problem in the natural
rubber industry. European Journal of Operational Research 6, 183-189.
Nambiar, J.M., Gelders, L.F., Van Wassenhove, L.N., 1989, Plant location and vehicle routing in the Malaysian
rubber smallholder sector: a case study. European Journal of Operational Research 38, 14-26.
Ogryczak, W., Studziński, K., Zorychta, K., 1989, A solver for the multi-objective transshipment problem with
facility location. European Journal of Operational Research 43, 53-64.
Ogryczak, W., Studziński, K., Zorychta, K., 1992, DINAS: a computer-assisted analysis system for multiobjective
transshipment problems with facility location. Computers and Operations Research 19, 637-647.
Or, I., Pierskalla, W.P., 1979, A transportation location-allocation model for regional blood banking, AIIE
Transactions 11, 86-94.
Perl, J., 1983, A Unified Warehouse Location-Routing Analysis, PhD dissertation, Northwestern University.
Perl, J., Daskin, M.S., 1984, A unified warehouse location-routing methodology. Journal of Business Logistics 5,
92-111.
Perl, J., Daskin, M.S., 1985, A warehouse location-routing problem. Transportation Research B 19, 381-396.
Rand, G.K., 1976, Methodological choices in depot location studies. Operational Research Quarterly 27, 241-249.
Reece, K., 1985, Vehicle Routing Considerations in Distribution System Design, M.A.Sc. thesis, University of
Waterloo.
ReVelle, C., Cohon, J., Shobrys, D., 1991, Simultaneous siting and routing in the disposal of hazardous wastes.
Transportation Science 25, 138-145.
Rodríguez-Martín, I., 2000, Cycle Location Problems, PhD dissertation, University of La Laguna.
Salhi, S., 1987, The Integration of Routing into the Location-Allocation and Vehicle Fleet Composition Problems,
PhD dissertation, Lancaster University.
Salhi, S., Fraser, M., 1996, An integrated heuristic approach for the combined location vehicle fleet mix problem.
Studies in Locational Analysis 8, 3-21.
Salhi, S., Nagy, G., 1999, Consistency and robustness in location-routing. Studies in Locational Analysis 13, 3-19.
Salhi, S., Nagy, G., 2007, Local improvement in planar facility location using vehicle routing. Annals of Operations
Research (in review).
26
Salhi, S., Rand, G.K., 1989, The effect of ignoring routes when locating depots. European Journal of Operational
Research 39, 150-156.
Schwardt, M., Dethloff, J., 2005, Solving a continuous location-routing problem by use of a self-organising map.
International Journal of Physical Distribution and Logistics Management 35, 390-408.
Semet, F., 1995, A two-phase algorithm for partial accessibility constrained vehicle routing problem. Annals of
Operations Research 61, 45-65.
Semet, F., Taillard, E., 1993, Solving real-life vehicle routing problems efficiently using tabu search. Annals of
Operations Research 41, 469-488.
Souid, M., 2003, Résolution Approchée d’un Problèm de Tournées de Véhicule avec Contraintes d’Accessibilité.
PhD thesis, University of Valenciennes.
Souid, M., Hanafi, S., Semet, F., 2007, Variable neighborhood search for the vehicle routing problem with
accessibility constraints. 4OR (in review)
Shen, Z-J.M., Coullard, C., Daskin, M.S., 2003, A joint location-inventory model. Transportation Science 37, 40-55.
Simchi-Levi, D., 1987, The Multistop Location Problems, PhD dissertation, University of Tel-Aviv.
Simchi-Levi, D., 1991, The capacitated traveling salesman location problem. Transportation Science 25, 9-18.
Simchi-Levi, D., Berman, O., 1987, Heuristics and bounds for the travelling salesman location problem on the plane.
Operations Research Letters 6, 243-248.
Simchi-Levi, D., Berman, O., 1988, A heuristic algorithm for the traveling salesman location problem on networks.
Operations Research 36, 478-484.
Srivastava, R., 1986, Algorithms for Solving the Location-Routing Problem, PhD dissertation, Ohio State
University.
Srivastava, R., 1993, Alternate solution procedures for the location-routing problem. Omega 21, 497-506.
Srivastava, R., Benton, W.C., 1990, The location-routing problem: considerations in physical distribution system
design. Computers in Operations Research 17, 427-435.
Stowers, C.L., Palekar, U.S., 1993, Location models with routing considerations for a single obnoxious facility.
Transportation Science 27, 350-362.
Tham, W.C., 2005, Depot location-routing models based on a real supply chain network. PhD dissertation, Lancaster
University.
Toth, P., Vigo, D., 2002a, The Vehicle Routing Problem. SIAM, Philadelphia.
Toth, P., Vigo, D., 2002b, Models, relaxations and exact approaches for the capacitated vehicle routing problem.
Discrete Applied Mathematics 123, 487-512.
Türkay, A., Emel, E., 2007, Vehicle routing problem with packing constraints. European Journal of Operational
Research (in review).
Tuzun, D., Burke, L.I., 1999, A two-phase tabu search approach to the location routing problem. European Journal
of Operational Research 116, 87-89.
Verter, V., Dincer, M.C., 1995, Facility location and capacity acquisition: an integrated approach. Naval Research
Logistics 42, 1141-1160.
Wasner, M., Zäpfel, G., 2004, An integrated multi-depot hub-location vehicle routing model for network planning of
parcel service. International Journal of Production Economics 90, 403-419.
Watson-Gandy, C.D.T., Dohrn, P.J., 1973, Depot location with van salesmen – a practical approach. Omega 1, 321329.
Wu, T-H., Low, C., Bai, J-W., 2002, Heuristic solutions to multi-depot location-routing problems. Computers and
Operations Research 29, 1393-1415.
Webb, M.H.J., 1968, Cost functions in the location of depots for multiple-delivery journeys. Operational Research
Quarterly 19, 311-320.
Zografos, K.G., Samara, S., 1989, A combined location-routing model for hazardous waste transportation and
disposal. Transportation Research Record 1245, 52-59.
27
http://www.kent.ac.uk/kbs/research-information/index.htm