Computers & Industrial Engineering 79 (2015) 115129

Contents lists available at ScienceDirect

Computers & Industrial Engineering

journal homepage: www.elsevier.com/locate/caie

Survey

A literature review on the vehicle routing problem with multiple depots

Jairo R. Montoya-Torres a,, Julin Lpez Franco b, Santiago Nieto Isaza c, Heriberto Felizzola Jimnez d,
Nilson Herazo-Padilla e,f
a
Escuela Internacional de Ciencias Econmicas y Administrativas, Universidad de La Sabana, km 7 autopista norte de Bogot D.C., Chia (Cundinamarca), Colombia
b
Engineering and Consulting SAS, Bogot, D.C., Colombia
c
Departamento de Ingeniera Industrial, Universidad del Norte, Km 5 va Puerto Colombia, Barranquilla (Atlntico), Colombia
d
Departamento de Ingeniera Industrial, Universidad de La Salle, Carrera 2 # 10-70, Bogot, D.C., Colombia
e
Departamento de Ingeniera Industrial, Universidad de la Costa, Calle 58 # 55-66, Barranquilla, Colombia
f
Fundacin Centro de Investigacin en Modelacin Empresarial del Caribe FCIMEC, Carrera 53 # 74-86 Oc.402, Barranquilla, Colombia

a r t i c l e i n f o a b s t r a c t

Article history: In this paper, we present a state-of-the-art survey on the vehicle routing problem with multiple depots
Received 3 August 2012 (MDVRP). Our review considered papers published between 1988 and 2014, in which several variants of
Received in revised form 28 October 2014 the model are studied: time windows, split delivery, heterogeneous eet, periodic deliveries, and pickup
Accepted 31 October 2014
and delivery. The review also classies the approaches according to the single or multiple objectives that
Available online 10 November 2014
are optimized. Some lines for further research are presented as well.
2014 Elsevier Ltd. All rights reserved.
Keywords:
Vehicle routing
Multiple depots
Exact algorithms
Heuristics
Survey

1. Introduction Gutirrez-Franco, & Halabi, 2009; Ozrat & Ozkarahan, 2010;

Thangiah & Salhi, 2001). Solving the VRP is vital in the design of
Physical distribution is one of the key functions in logistics distribution systems in supply chain management.
systems, involving the ow of products from manufacturing plants
or distribution centers through the transportation network to
consumers. It is a very costly function, especially for the distribu- 1.1. VRP versus MDVRP
tion industries. The Operational Research literature has addressed
this problem by calling it as the vehicle routing problem (VRP). The Formally, the classical vehicle routing problem (VRP) is repre-
VRP is a generic name referring to a class of combinatorial optimi- sented by a directed graph G(E,V), where V = {0,1, . . ., n} represents
zation problems in which customers are to be served by a number the set of nodes and E is the set of arcs. The depot is noted to be
of vehicles. The vehicles leave the depot, serve customers in the node j = 0, and clients are nodes j = 1, 2, . . ., n, each one with
network and return to the depot after completion of their routes. demand dj > 0. Each arc represents a route from node i to node j.
Each customer is described by a certain demand. This problem The weight of each arc Cij > 0 corresponds to the cost (time or even
was rstly proposed in the literature by Dantzig and Ramser distance) of going from node i to node j. If Cij = Cji then we are
(1959). After then, considerable number of variants has been facing the symmetric VRP, otherwise the problem is asymmetric.
considered: hard, soft and fuzzy service time windows, maximum From the complexity point of view, the classical VRP is known to
route length, pickup and delivery, backhauls, etc. (Cordeau, NP-hard since it generalizes the Travelling Salesman Problem
Gendreau, Hertz, Laporte, & Sormany, 2005; Juan, Fauln, (TSP) and the Bin Packing Problem (BPP) which are both
Adelanteado, Grasman, & Montoya Torres, 2009; Lpez-Castro & well-known NP-hard problems (Garey & Johnson, 1979). A review
Montoya-Torres, 2011; Montoya-Torres, Alfonso-Lizarazo, of mathematical formulations for the classical VRP can be found in
the work of Laporte (1992).
Corresponding author. In the literature, lots of surveys have been presented analyzing
E-mail addresses: [email protected] (J.R. Montoya-Torres), julian.
published works on either the classical version (Bodin, 1975; Bodin
[email protected] (J. Lpez Franco), [email protected] (S. Nieto Isaza), & Golden, 1981; Desrochers, Lenstra, & Savelsbergh, 1990;
[email protected] (H. Felizzola Jimnez). Eksioglu, Volkan, & Reisman, 2009; Laporte, 1992; Liong, Wan,

http://dx.doi.org/10.1016/j.cie.2014.10.029
0360-8352/ 2014 Elsevier Ltd. All rights reserved.
116 J.R. Montoya-Torres et al. / Computers & Industrial Engineering 79 (2015) 115129

Khairuddin, & Mourad, 2008; Mafoli, 2002) or its different vari-

ants: the capacitated VRP (Baldacci, Toth, & Vigo, 2010; Cordeau,
Laporte, Savelsbergh, & Vigo, 2007; Gendreau, Laporte, & Potvin,
2002; Laporte & Nobert, 1987; Laporte & Semet, 2002; Toth &
Vigo, 2002), the VRP with heterogeneous eet of vehicles
(Baldacci, Battarra, & Vigo, 2008; Baldacci, Toth, & Vigo, 2007;
Baldacci et al., 2010), VRP with time windows (VRPTW), pickup
and deliveries and periodic VRP (Solomon & Desrosiers, 1988),
dynamic VRP (DVRP) (Psaraftis, 1995), Periodic VRP (PVRP)
(Mourgaya & Vanderbeck, 2006), VRP with multiple trips (VPRMT)
(S
en & Blbl, 2008), Split Delivery vehicle routing problem
(SDVRP) (Archetti & Speranza, 2008). All of these works consider
only one depot. Fig. 1 presents a hierarchy of VRP variants. One
of these variants considers a well-known (Crevier, Cordeau, &
Laporte, 2007) more realistic situation in which the distributions
of goods is done from several depots to nal clients. This particular
distribution network can be solved as multiple individual single
depot VRPs, if and only if clients are evidently clustered around
each depot; otherwise a multi-depot-based approach has to be
used where clients are to be served from any of the depots using
the available eet of vehicles. In this paper, we consider the variant Fig. 2. Comparison between VRP versus MDVRP.
of the vehicle routing problem known as Multiple Depots Vehicle
Routing Problem (MDVRP) in which more than one depot is consid-
ered (see Fig. 2). The reader must note that most exact algorithms demand of each route does not exceed the vehicle capacity, and
for solving the classical VRP model are difcult to be adapted for (iv) the total cost of the distribution is minimized. According to
solving the MDVRP. Kulkarni and Bhave (1985), a mathematical model of the MDVRP
According to Renaud, Laporte, and Boctor (1996), the MDVRP requires the denition of binary decision variables xijk to be equal
can be formally described as follows. Let G = (V,E) be a graph, to 1 if the pair of nodes i and j are in the route of vehicle k, and
where V is the set of nodes and E is the set of arcs or edges 0 otherwise. Auxiliary variables yi are required in order to avoid
connecting each pair of nodes. The set V is further partitioned into subtour elimination. According to this last reference, the model is
two subsets: Vc = {v1, v2, . . . , vN} which is the set of customers to be as follows:
served; and Vd = {vN+1, vN+2, . . . , vM} which is the set of depots. Each
customer vi 2 Vc has a nonnegative demand di. Each arc belong to
X NM
NM XXK
the set E has associated a cost, distance or travel time cij. There Minimize cij xijk 1
are a total of K vehicles, each one with capacity Pk. The problem i1 j1 k1
consists on determining a set of vehicle routes in such a way that: Subject to :
(i) each vehicle route starts and ends at the same depot, (ii) each
XX
NM K
customer is serviced exactly once by a vehicle, (iii) the total xijk 1 j 1; .. .; N 2
i1 k1
XX
NM K
xijk 1 i 1;. .. ;N 3
j1 k1

periodic Multiple X
N M X
N M
k 1;. .. ;K
VRP depots xihk  xhjk 0 4
i1 j1
h 1; .. .; N M
constraints

distance or time
X
NM X
NM
capacity

constraints time windows Qi xijk 6 Pk k 1; .. .; K 5

i1 j1

X NM
NM X
cij xijk 6 T k k 1; .. .; K 6
i1 j1
capacity & time of
X
NM X
N
with loading
& unloading

distance constraints
backhaul xijk 6 1 k 1; .. .; K 7
iN1 j1

X
NM X
N
xijk 6 1 k 1; .. .; K 8
jN1 i1

yi  yj M Nxijk 6 N M  1 For 1 6 i j 6 N and 1 6 k 6 K 9

depot is source
& destination

xijk 2 f0; 1g 8i; j;k 10

In this formulation, Constraints (2) and (3) ensure that each
customer is served by one and only one vehicle. Route continuity
is represented by Constraints (4). The sets of constraints (5) and
(6) are the vehicle capacity and total route cost constraints. Vehicle
availability is veried by Constraints (7) and (8) and subtour elim-
ination is provided by Constraints (9). In this formulation, it is
Fig. 1. Different variants of the VRP (adapted from Weise et al., 2010). assumed that total demand at each node is either less than or at
 J.R. Montoya-Torres et al. / Computers & Industrial Engineering 79 (2015) 115129 117

the most equal to the capacity of each vehicle. Similar to single short-listed in this review (see Tables A1 and A2 in Appendix).
depot VRP, the subtour elimination constraint (9) can be rewritten The short-listed publications are then examined in more detail.
in a more compact form as: The analysis of methodological issues and problem variants are
presented in detail in Section 4 of the paper.
X
V
yi  yj M N xijk 6 M N  1 for i 6 ij 6 M N  1
k1 1.4. Paper organization
11
The rest of this paper is organized as follows. Section 2 presents
the review of papers studying the single-objective problem, while
1.2. Motivation for a review of research papers on MDVRP those focused on multiple objectives are reviewed in Section 3. In
both sections, papers are briey commented and most important
The MDVRP is more challenging and sophisticated than the sin- facts are highlighted. The literature is quantied and measured in
gle-depot VRP. The variant with multiple depots appears rst in Section 4. The paper ends in Section 5 by presenting some conclud-
the literature on the works of Kulkarni and Bhave (1985), and ing remarks and discussing some research opportunities.
Laporte, Nobert, and Taillefer (1988) and Carpaneto, Dellamico,
Fischetti, and Toth (1989). Since then, considerable amount of
2. Single objective MDVRP
research has been published (see Table 1) in the form of journal
paper, conference paper, research/technical report, thesis or book.
This section presents a literature review of papers considering
To the best of our knowledge, despite the great amount of research
the single objective case of the MDVRP. For the total of papers con-
papers published, there is no rigorous literature survey exclusively
sidered in this review, 88.44% considers only one optimization
devoted to the vehicle routing problem with multiple depots. A
objective. The list of reviewed papers is presented in Table A1 in
short overview of academic works was proposed by Liu, Jiang,
the Appendix. This section is divided into three main parts, each
Liu, and Liu (2011), but only presenting the most representative
one corresponding to the type of solution procedure employed:
research papers. From the 58 cited references in their paper, only
exact method, heuristic or meta-heuristic. After the rst works
23 of them explicitly refer to the MDVRP. Besides, these authors
were published in the decade of 1980, more than one hundred of
focus on the problem denition, solution methods (dividing them
papers have studied the classical version of the MDVRP and its
into exact algorithms, heuristics and meta-heuristics) and mention
variants, some of them inspired from real-life applications.
some problem variants. In fact, no actual systematic review was
presented in that paper. Our intention now is to present a rigorous
review of scientic literature, by presenting a taxonomic classica- 2.1. Exact methods
tion of those works. Most of the published works focus on the sin-
gle objective problem, while a few consider the multi-objective In the decade of the 1970s, some works already mentioned
case. In this paper, we intend to provide an analysis of both single some problems related to distribution of goods with multiple
and multiple objective problems. depots. However, to the best of our knowledge, the rst paper pre-
senting formal models or procedures to nd optimal solution for
the multi-depot vehicle routing problem are those of Laporte,
1.3. Review methodology
Nobert, and Arpin (1894) who formulated the symmetric MDVRP
as integer linear programs with three constraints. These authors
This paper presents a review of relevant literature on the
then proposed a branch-and-bound algorithm using a LP relaxa-
vehicle routing problem with multiple depots, with both single
tion. The works of Kulkarni and Bhave (1985), Laporte et al.
and multiple objective functions. An ambitious search was con-
(1988) and Carpaneto, Dellamico, Fischetti, and Toth (1989) can
ducted using the library databases covering most of the major jour-
also be considered as part of the pioneer works on exact methods
nals. Some conference papers are also included in this review. In
for the MDVRP. The mathematical formulation proposed by
addition, the websites of leading research groups and the principal
Kulkarni and Bhave (1985) was later revised by Laporte (1989).
authors of major publications were also searched for further infor-
More recently, Baldacci and Mingozzi (2009) proposed mathemat-
mation about their research projects (Ph.D. projects and sponsored
ical formulations for solving several classes of vehicle routing
projects) and publications. We intentionally excluded working
problems including the MDVRP, while Nieto Isaza, Lpez Franco,
papers, theses and research reports not available online on the
and Herazo Padilla (2012) presented an integer linear program
Internet because they are very difcult to obtain.
for solving the heterogeneous eet MDVRP with time windows.
The initial collection of references was screened rst for their
Dondo, Mendez, and Cerd (2003) proposed a mixed-integer linear
relevance and their signicance for the purpose of this review. In
programming (MILP) model to minimize routing cost in the
order to control the length of this paper, only some representative
HFMDVRP, in which heterogeneous eet of vehicles are available.
publications were selected to be explained in detail within the text
The variant with pickup and deliveries and heterogeneous eet
of this manuscript, which are authored by leading researchers or
was modeled by Dondo, Mndez, and Cerd (2008) using MILP
groups. These selected authors and research groups have, in fact,
model: this model is able to solve only small sized instances, and
published a long list of research papers and reports in the eld. A
hence a local search improvement algorithm was then proposed
collection of a total of 147 representative publications are
by the authors for medium to large sized instances. This approach
Table 1 was later employed by Dondo and Cerd (2009) to solve the
Number and types of publications on MDVRP. HFMDVRPTW. The work of Kek, Cheu, and Meng (2008) proposes
Type of publication Total a mixed-integer linear programming model and a branch-and-
Journal 125
bound procedure for the MDVRP with xed eet and pickup and
Conference 28 delivery. The objective function is the minimization of the total
Thesis 7 cost of routes. Cornillier, Boctor, and Renaud (2012) presented a
Technical report 4 MILP model for the problem in which heterogeneous eet of
Book/book chapter 9
vehicles is available and with maximization of total net revenue
Total 173
as objective function, while maximum and minimum demands
118 J.R. Montoya-Torres et al. / Computers & Industrial Engineering 79 (2015) 115129

constraints are given. Seth, Klabjan, and Ferreira (2013) studied an experimental results showed that the one-stage algorithm outper-
application case from the printed circuit board production process forms the other one.
modeled as a MDVRP with mobile depots. They presented an exact The HFMDVRP, in which heterogeneous eet of vehicles is
algorithm based on network-ow formulation and proposed a iter- available have captured the attention of researchers since the work
ated tour partitioning heuristic. Branch-and-cut algorithms were presented by Salhi and Sari (1997). Irnich (2000) proposed a set
proposed by Benavent and Martnez (2013) and Braekers, Caris, covering heuristic coupled with column generation and branch-
and Jenssens (2014). Former authors focused also on studying and-price algorithm for cost minimization for the heterogeneous
the polyhedral structure of the problem which allowed the exten- eet and pickup and delivery MDVRP. Dondo and Cerd (2007)
sion of their procedure to the location-routing problem (LRP), proposed a MILP model as well as a three-stage heuristic. Before,
while latter authors considered a dial-a-ride problem with multi- a preprocessing stage for node clustering is performed and a more
ple depots. The LRP with multiple depots was also studied by compact cluster-based MILP problem formulation is developed.
Contardo, Cordeau, and Gendron (2014) who proposed a cut-and- Many other papers have been appeared in literature on or before
column generation procedure for the capacitated case. Other 2007 and solution approaches have primarily been focused on
congurations of the MDVRP have been studied through exact meta-heuristic algorithms. Hence, this will be discussed more in
algorithms. For instance, Contardo and Martinelli (2014) studied detail in the next subsection.
the capacitated MDVRP with route length constraints; Ray, Concerning the periodic MDVRP, few works appear in literature
Soeanu, Berger, and Debbabi (in press) modeled the version with with heuristic algorithm as solution approach. We have identied
split deliveries; Li, Li, and Pardalos (in press) proposed an integer only the works of Hadjiconstantinou and Baldacci (1998), Vianna,
programming model for the MDVRP with shared depots and time Ochi, and Drummond (1999), Yang and Chu (2000), Maischberger
windows; and Adelzadeh, Asi, and Koosha (in press) proposed a and Cordeau (2011), and Maya, Srensen, and Goos (2012). The
mathematical model and a fuzzy-based heuristic for the particular problem with time windows was studied by Chiu, Lee, and Chang
case of the MDVRP with fuzzy time windows and heterogeneous (2006) who presented a two-phase heuristic method. In contrast
eet of vehicles. with other works in literature, these authors considered the wait-
ing time as objective function. Results indicate that the waiting
time has a signicant impact on the total distribution time and
2.2. Heuristics the number of vehicles used when solving test problems with nar-
row time windows. The authors also considered a real-life case
Because the NP-hardness of the MDVRP, several heuristic study of a logistics company in Taiwan.
algorithms have been proposed in the literature. This section sum- Tsirimpas, Tatarakis, Minis, and Kyriakidis (2007) considered
marizes some of the most relevant works concerning different vari- the case of a single vehicle with limited capacity, multiple-prod-
ants of the problem. The rst works were published in the 1990s, ucts and multiple depot returns. Another characteristic of their
in order to solve the capacitated version. Min, Current, and problem is that the sequence of visits to customer is predened.
Schilling (1992) studied the version of the MDVRP with backhaul- They developed a suitable dynamic programming algorithm for
ing and proposed a heuristic procedure based on problem the determination of the optimal routing policy. For the MDSDVRP
decomposition. Hadjiconstantinou and Baldacci (1998) considered which consists on the combination of the MDVRP and the Split
a real-life problem taken from a company supplying maintenance Delivery VRP (SDVRP). The work of Gulczynski, Golden, and
services. Their problem consists on determining the boundaries Wasil (2011) developed an integer programming-based heuristic.
of the geographic areas served by each depot, the list of customers The objective of this study was to determine the reduction in trav-
visited each day and the routes followed by the gangs. The objec- eled distance that can be achieved by allowing split deliveries
tive is to provide improved customer service at minimum operat- among vehicles based at the same depot and vehicles based at dif-
ing cost subject to constraints on frequency of visits, service time ferent depots. The multi-depot capacitated vehicle routing prob-
requirements, customer preferences for visiting on particular days lem with split delivery (MDCVRPSD) is studied by Liu, Jiang,
and other routing constraints. This situation was solved using the Fung, Chen, and Liu (2010). A mathematical model is proposed
periodic variant: MDPVRP for which a ve-level heuristic was pro- based on a graph model. The objective function is the minimization
posed: rst and second levels solve the problem of determining the of movements of empty vehicle. A greedy algorithm is proposed as
service areas and service days (the periodic VRP); third level solves well, in order to solve large-scale instances. More recent works on
the VRP for each day; fourth level solves a TSP for each route, and the application of dedicated heuristics include the work of Vahed,
fth level seeks the optimization of routes. Crainic, Gendreau, and Rei (in press) for the case of a MDVRP with
Salhi and Sari (1997) proposed the so-called multi-level com- the objective of determining the optimal vehicle eet size, and the
posite heuristic. This heuristic found as good solutions as those work of Afshar-Nadja and Afshar-Nadja (in press) for the study
known at that time in the literature but using only 5 to 10% of their of the time-dependent MDVRP with heterogeneous vehicles and
computing time. The heuristic was also tested on the problem with time windows.
heterogeneous eet. Salhi and Nagy (1999) proposed an insertion-
based heuristic in order to minimize routing cost. Later, these 2.3. Meta-heuristics
authors (Nagy & Salhi, 2005) also studied the problem with pickup
and deliveries (MDVRPPD). Their approach avoids the concept of As for other NP-hard combinatorial optimization problems,
insertion and proposes a method that treats pickups and deliveries meta-heuristic procedures have been employed by several
in an integrated manner. The procedure rst nds a solution to the researchers for efciently solving the single-objective MDVRP.
VRP, then it modies this solution to make it feasible for the VRPPD The rst meta-heuristic was proposed in the work of Renaud,
and it nally ensures that it is also feasible for the multi-depot Laporte et al. (1996) who study the MDVRP with the constraints
case. Jin, Guo, Wang, and Lim (2004) modeled the MDVRP as a of vehicle capacities and maximum duration of routes (e.g. the
binary programming problem. Two solving methodologies were time of a route cannot exceed the maximum working time of the
presented. The rst one is a two-stage approach that decomposes vehicle). The objective to be optimized is the total operational cost.
and solves the problem into two independent subproblems: These authors proposed a Tabu Search algorithm for which the ini-
assignment and routing. The second proposed approach treats both tial solution is built using the Improved Petal heuristic of Renaud,
assignment and routing problems in an integrated manner. Their Boctor, and Laporte (1996b). Experiments were carried out using
 J.R. Montoya-Torres et al. / Computers & Industrial Engineering 79 (2015) 115129 119

classical instances of Christodes and Eilon (1969) and Gillett and reverse logistics problem in which the aim is to collect cores from
Johnson (1976). Later, Cordeau, Gendreau, and Laporte (1997) an enterprises dealers. The problem is called the selective MDVRP
proposed a Tabu Search algorithm with the initial solution being with price. In addition to the Tabu Search procedure, the authors
generated randomly for the MDVRP that can also be used to solve also formulated two mixed-integer linear programming (MILP)
the periodic VRP (PVRP), while Tzn and Burke (1999) proposed a models.
Tabu Search procedure for minimizing the total cost of the routing. Other meta-heuristics, such as GRASP, are presented in the
Cordeau, Laporte, and Mercier (2001) also proposed a TS procedure works of Villegas, Prins, Prodhon, Medaglia, and Velasco (2010)
with the objective of minimizing the number of vehicles. An and Maya et al. (2012), respectively minimizing route cost and
approximation to real industrial practice was studied by distance. zyurt and Aksen (2007) solved the problem of depot
Parthanadee and Logendran (2006). In their problem, depots location and vehicle routing using a hybrid approach based on
operate independently and solely within their own territories. Lagrangian relaxation (LR) and Tabu Search (TS). These procedures
The distributors may hence satisfy customers requests by deliver- improve the best solutions found for the set of instances proposed
ing products from other depots that hold more supplies. They by Tzn and Burke (1999). A case study taken from waste collec-
proposed a mixed-integer linear programming model for the tion system involving 202 localities in the city of Viseu, Portugal, is
multi-product, multi-depot periodic distribution problem and pre- presented by Matos and Oliveira (2004). An Ant Colony Optimiza-
sented three Tabu Search heuristics with different long-term mem- tion (ACO) algorithm is proposed and compared with other proce-
ory applications. Results revealed that interdependent operations dures from the literature.
provide signicant savings in costs over independent operations The great amount of heuristics algorithms proposed for the
among depots, especially for large-size problems. More recently, problem variant with heterogeneous eet (HFMDVRP) has been
Escobar, Linfati, Toth, and Baldoquin (in press) evaluated a Granu- focused on the design of meta-heuristics algorithms. We can
lar Tabu Search algorithm to minimize the total sum of vehicle highlight the works of Jeon, Leep, and Shim (2007), who proposed
traveled distances. a hybrid genetic algorithm that minimizes the total distance
The rst genetic algorithms were proposed by Filipec, Skrlec, traveled, and that of Flisberg, Lidn, and Rnnqvist (2009) who
and Krajcar (1997) for the problem of minimizing total travel dis- considered a Tabu Search procedure. Simulated Annealing (SA)
tance, by Salhi, Thangiah, and Rahman (1998) and by Skok, Skrlec, has been employed as well. Wu et al. (2002) coupled SA with Tabu
and Krajcar (2000), Skok, Skrlec, and Krajcar (2001). An evolution- Search to solve the heterogeneous eet case of the integrated loca-
ary algorithm coupled with local search heuristic was proposed by tion-routing problem. In their problem, location of depots, routes
Vianna et al. (1999) in order to minimize the total cost. Thangiah of vehicles and client assignment problems are solved simulta-
and Salhi (2001) proposed the use of genetic algorithms to dene neously. The multi-depot heterogeneous vehicle routing problem
clusters of clients and then routes are found by solving a traveling with time windows (MDHVRPTW) was studied by Dondo and
salesman problem (TSP) using and insertion heuristic. This Cerd (2009), who proposed a MILP and a Local Search Improve-
approach is called Genetic Clustering (GenClust). Solutions are ment Algorithm that explores the neighborhood in order to nd
nally optimized using the post-optimization procedure of Salhi the lowest cost feasible solution.
and Sari (1997). Recently, Ycenur and Demirel (2011a) proposed Other research papers have also been very interested in the
geometric shape based genetic clustering algorithm for the classi- analysis of the problem with time windows (MDVRPTW). This
cal MDVRP. The procedure is compared with the nearest neighbor variant is studied in 25% of the single-objective focused papers
algorithm. Their experiments showed that their algorithm provides considered in this review. The rst meta-heuristics reported in lit-
a better clustering performance in terms of the distance of each erature was the Tabu Search procedure of Cordeau et al. (2001) in
customer to each depot in clusters, in a considerably less computa- which the objective function is the minimization of the number of
tional time. vehicles. Polacek, Hartl, Doerner, and Reimann (2005) proposed a
In the survey by Gendreau, Potvin, Brumlaysy, Hasle, and Variable Neighborhood Search (VNS) algorithm for the MDVRP
Lkketangen (2008) focused on the application of meta-heuristics with time windows and with xed distribution of vehicles. This
for solving various variants of the VRP, a short revision of the problem was also studied by Lim and Wang (2005) with the char-
multi-depot problem is presented. The equivalence between the acteristic of having exactly one vehicle at each depot. Jin, Mitsuo,
MDVRP and the PVRP is also analyzed. Among the meta-heuristics and Hiroshi (2005), Yang (2008), Ghoseiri and Ghannadpour
presented therein, we can highlight the use of Genetic Algorithms (2010) and Samanta and Jha (2011) proposed Genetic Algorithms,
(Filipec, Skrlec, & Krajcar, 2000), Simulated Annealing (Lim & Zhu, Pisinger and Ropke (2007) presented an Adaptive Large Neighbor-
2006) of the case of xed vehicle eet, and Tabu Search (Angelelli hood Search (ALNS) procedure with minimization of routing cost.
and Speranza, 2002). Other works proposing meta-heuristics can Ting and Chen (2009) presented a hybrid algorithm that combines
be found in (Chao, Golden, & Wasil, 1993 and Chen, Takama, & multiple ant colony systems (ACS) and Simulated Annealing (SA).
Hirota, 2000). Zarandi, Hemmati, and Davari (2011) presented a SA procedure
The most studied variant of the problem has been the capaci- to minimize routing cost, while Wang, Zhang, and Wang (2011)
tated MDVRP. Among the meta-heuristics proposed in literature, coupled SA with a modied Variable Neighborhood Search algo-
we can highlight the Simulated Annealing algorithms of Wu, rithm, and a clustering algorithm is used to allocate customers in
Low, and Bai (2002) and Lim and Zhu (2006), the Variable Neigh- the initial solution construction phase. A branch-and-cut-and-
borhood Search procedure proposed by Polacek, Hartl, Doerner, price algorithm for the multi-depot heterogeneous vehicle routing
and Reimann (2005), Polacek, Benkner, Doerner, and Hartl problem with time windows (MDHVRPTW) was recently proposed
(2008), Tabu Search algorithms from Lim and Wang (2005), Aras, by Bettinelli, Ceselli, and Righini (2011). Computational experi-
Aksen, and Tekin (2011) and Maischberger and Cordeau (2011). ments showed that this procedure is competitive in comparison
Genetic Algorithms has been proposed as well for this problem with local search heuristics.
variant, as illustrated in the works of Bae, Hwang, Cho, and Goan The variants with split delivery (MDVRPSD) or with pickup &
(2007), Vidal, Crainic, Gendreau, Lahrichi, and Rei (2010). All of delivery (MDVRPPD) have been considered by very few authors
these works seek for the minimization of total route distance or in the scientic literature. The work of Wasner and Zapfel (2004)
cost, except the work of Aras et al. (2011) in which the objective presents an application to postal, parcel and piece goods service
is the maximization of vehicle utilization rate. It is to note here that provider in Austria. The model employed is the MDVRPPD (MDVRP
the work of Aras et al. (2011) is inspired by a particular case of with pickup and deliveries) with the objective of determining the
120 J.R. Montoya-Torres et al. / Computers & Industrial Engineering 79 (2015) 115129

number and location of depots and hubs. Also, the client assign- combines the exploration breadth of population-based evolution-
ment problem is addressed. These authors develop a local search ary search, the aggressive improvement capabilities of neighbor-
heuristic. As a real-life problem is solved, additional features are hood search based procedures and advanced population diversity
included in the algorithm in order to take into account the topog- management strategies. These authors improved the best-known
raphy of the country (which is characterized by mountains) by solutions and even obtained optimal values for these three problem
considering maximum route length. The decision support system cases. Recent years on MDVRP research have witnessed the use of
allowed the solution of large-sized instances with various millions variable neighborhood search (VNS) algorithm for the resolution
of variables and constraints. The paper of Pisinger and Ropke of various variants of the problem (Kuo & Wang, 2012; Salhi,
(2007) studied the MDVRPPD, together with the variants of time Imran, & Wassan, 2014; Xu & Jiang, 2014; Xu, Wang, & Yang,
windows and vehicle capacity constraint. These authors proposed 2012). A realistic application of MDVRP found in vessel routing
an Adaptive Large Neighborhood Search procedure in order to was studied by Hirsch, Schroeder, Maggiar, and Dolinskaya
minimize total routing cost. Flisberg, Lidn, and Rnnqvist (2009) (2014). These authors proposed the implementation of various heu-
also considered heterogeneous eet of vehicles and time windows ristics, including GRASP (Greedy Randomized Adaptive Search Pro-
constraints, in addition to pickups and split deliveries: their case- cedure). More recently, the trend has been focused on the use of
study is taken from a forest company in Sweden. Schmid, hybrid meta-heuristics algorithms. Liu and Yu (2013) presented a
Doerner, Hartl, and Salazar-Gonzlez (2010) studied a realistic case hybridized genetic algorithm ant colony optimization procedure
inspired from companies in the concrete industry, and presented a to minimize the maximum travel distance of a vehicle in a system
mixed integer linear program (MILP) and a Variable Neighborhood with heterogeneous eet of vehicles. Liu (2013) proposed to couple
Search (VNS) procedure to minimize routing cost for the variant the genetic algorithm with bee colony optimization and simulated
with split deliveries. Mirabi, Fatemi Ghomi, and Jolai (2010) annealing to solve the classical MDVRP. Rahimi-Vahed, Crainic,
addressed the problem of minimizing the delivery time of vehicles. Gendreau, and Rei (2013) employed path relinking for the case of
They compared three hybrid heuristics, each one combining capacitated MDVRP with route duration constraint. Vidal, Crainic,
elements from both constructive heuristic search and improve- Gendreau, and Prins (2014) proposed a hybrid genetic algorithm
ment techniques. The improvement techniques are deterministic, with iterated local search and dynamic programming was pre-
stochastic and simulated annealing (SA) methods. sented for the classical MDVRP with unconstrained vehicle eet.
Crevier et al. (2007) considered a MDVRP in which there are Subramanian, Uchoa, and Ochi (2013) proposed a matheuristic pro-
intermediate depots along vehicles routes where they may be cedure for the cases of the MDVRP and MDRVP with mixed pick up
replenished. This problem was inspired from a real-life application and deliveries. Their algorithm is based on iterated local search and
at the city of Montreal, Canada. A heuristic combining adaptive exploits set partitioning models at certain stages of the procedure
memory, tabu search and integer programming was proposed. to obtain competitive solutions. Sitek, Wikarek, and Grzybowska
The model allows the assignment of vehicles to routes that may (2014) presented a multi-agent system coupled with a mixed-inte-
begin and nish at the same depot or that connect two depots to ger linear programming (MILP) model and Constraint Programming
increase the capacity of vehicles to deliver goods. Zhen and (CP) for the multi-echelon capacitated vehicle routing problem.
Zhang (2009) considered a similar problem and proposed a heuris-
tic combining the adaptive memory principle, a Tabu Search
method for the solution of subproblems, and integer programming. 3. The MDVRP with multiple objectives
Another variant of the MDVRP appears in the work of Zarandi et al.
(2011). These authors studied the fuzzy version of the Capacitated An important characteristic of real-life logistics problems found
Location-Routing Problem (CLRP) with multiple depots in which in enterprises is that decision-makers, very often, have to simulta-
the location of depots have to be dened as well as the routes of neously deal with multiple objectives. These objectives are some-
vehicles. Fuzzy travel times between nodes and time window to times contradictory (e.g., minimizing number of vehicles and
meet the demand of each customer are considered. A simulation- maximizing service level). In the literature, there are very few
embedded Simulated Annealing (SA) procedure was proposed. papers on the MDVRP that consider multiple objectives (MOM-
The framework was tested using standard data sets. DVRP): about 11.56% of the papers reviewed here. This percentage
A good manner of improving the performance of meta-heuris- corresponds to a total of 17 papers reviewed here (see Table A2 in
tics is to generate good initial solutions. Ho, Ho, Ji, and Lau the Appendix).
(2008) proposed the use of the well-known Clarke & Wright The work of Lin and Kwok (2005) studies a realistic particular
Savings (C&WS) algorithm (Clarke & Wright, 1964) to generate case of the MDVRP, named as location-routing problem (LRP) with
initial solutions, as commonly used for other vehicle routing prob- multiple uses of vehicles. In this problem, decisions on location of
lems (Juan et al., 2009). Once the solution is generated, the Nearest depots, vehicle routing and assignment of routes to vehicles are
Neighbor (NN) heuristic is employed to improve such solution. In considered simultaneously. Tabu search and simulated annealing
comparison with the random generation of initial solutions, their procedures are designed and tested on both random-generated
experiments showed that this hybrid C&WS + NN approach pro- and real data. The objectives are the minimization of total opera-
duces better results regarding total delivery time. Li and Liu tional cost and the balance on vehicle load. Both simultaneous
(2008) considered the multi-depot open vehicle routing problem and sequential procedures for the assignment of routes to vehicles
with replenishment during the execution of routes. They proposed are tested. Results show that the simultaneous versions have
a model and an Ant Colony Optimization resolution procedure. advantage over the sequential versions in problems where routes
Other application of the Ant Colony Optimization paradigm can are capacity-constrained, but not in the time dimension. The
be found in the works of Wang (2013) and Narasimha, simultaneous versions are also more effective in generating non-
Kivelevitch, Sharma, and Kumar (2013). These last authors studied dominated solutions than the sequential versions.
the MDVRP with minimization of the longest travel distance of a A real-life transportation problem consisting on moving empty
vehicle. or laden containers is studied by Tan, Chew, and Lee (2006). They
Vidal et al. (2010, 2011) proposed a general framework to solve a called the problem as the truck and trailer vehicle routing problem
family of vehicle routing problems, including the multi-depot VRP, (TTVRP), but in fact it corresponds to a variant of the MDVRP: the
the periodic VRP and the multi-depot periodic VRP with capacitated solution to the TTVRP consists of nding a complete routing
vehicles and constrained route duration. Their meta-heuristic schedule for serving the jobs with minimum routing distance
 J.R. Montoya-Torres et al. / Computers & Industrial Engineering 79 (2015) 115129 121

160

140

120

100
Number of
80 papers

60 52 51

40
22
20 13
2 4 3
0
1980-1985 1986-1990 1991-1995 1996-2000 2001-2005 2006-2010 2011-2014

Fig. 3. Distribution of published papers per year for the MDVRP.

Fig. 4. Distribution of objective functions.

well-known algorithms non-dominated sorting genetic algorithms

2 (NSGA2), strength Pareto evolutionary algorithm 2 (SPEA2) with
and without fuzzy logic and the micro-genetic algorithm (MIC-
ROGA) with and without fuzzy logic. Experiments showed that
the proposed FL-NSGA2 procedure outperformed the other
procedures. This technique was also used by Lau, Chan, Tsui, and
Pang (2010) to solve the problem in which the cost due to the total
traveling distance and the cost due to the total traveling time are
minimized. In their work, several search methods, branch-and-
bound, standard GA (i.e., without the guide of fuzzy logic),
simulated annealing, and tabu search procedure are adopted to
Fig. 5. Distribution of employed solution techniques. compare with FLGA in randomly generated data sets. Results of
their experiments show that FLGA outperforms the other methods.
Ombuki-Berman and Hanshar (2009) and Weise, Podlich, and
and number of trucks, subject to time windows and availability of Gorldt (2010) also proposed a genetic algorithm. The rst authors
trailers. These authors solved the multi-objective case using a considered the objectives of minimizing the total cost and the
hybrid multi-objective evolutionary algorithm (HMOEA) with number of vehicles, while the latter authors considered the total
specialized genetic operators, variable-length representation and distance, the idle capacity of vehicles and the number of
local search heuristic. Lau, Tang, Ho, and Chan (2009) considered externalized deliveries. A Simulated Annealing (SA) procedure
the multi-objective problem in which the travel time is not a con- was presented by Hasanpour, Mosadegh-Khah, and Tavakoli
straint but an objective function to be optimized in the model Moghadam (2009) for minimizing transportation costs and
together with the total traveled distance. The proposed solution maximizing probability of delivery to customers.
procedure is a hybrid meta-heuristic named fuzzy logic guided Weise, Podlich, and Gorldt (2010) presented the use of
non-dominated sorting genetic algorithm (FL-NSGA2). The proce- evolutionary computation for a real-life problem inspired from a
dure uses fuzzy logic to dynamically adjust the probabilities of joint enterprise-academia research project. Results of the imple-
mutation and crossover. The algorithm is compared with the mentation are compared against the traditional routing structure
Table A1

122
Reviewed papers about the single-objective MDVRP.

Year Reference Constraints Solution method

Time Windows Heterogeneous eet Capacitated Periodic Pick up & Delievery Split delivery Other Exact Heuristic Meta-heuristic
% of papers 26% 21% 38% 12% 11% 5% 26% 31% 45% 57%
1984 Laporte et al. (1894) X
1985 Kulkarni and Bhave (1985) X X
1987 Laporte and Nobert (1987) X
1988 Laporte et al. (1988) X X
1989 Laporte (1989) X
Carpaneto, Dellamico, Fischetti, and Toth (1989) X X
1992 Min et al. (1992) X X X
Wilger and Maurer (1992) X
1993 Chao et al. (1993) X
1996 Renaud, Laporte et al. (1996) X X X

J.R. Montoya-Torres et al. / Computers & Industrial Engineering 79 (2015) 115129

1997 Filipec et al. (1997) X X
Salhi and Sari (1997) X X
Cordeau et al. (1997) X X
1998 Salhi et al. (1998) X
Hadjiconstantinou and Baldacci (1998) X X
1999 Vianna et al. (1999) X X X
Tzn and Burke (1999) X X
Salhi and Nagy (1999) X X
2000 Irnich (2000) X X X X
Filipec et al. (2000) X X X
Skok, Skrlec, and Krajcar (2000) X X X
Yang and Chu (2000) X X
2001 Thangiah and Salhi (2001) X X
Skok, Skrlec, and Krajcar (2001) X X X
Cordeau et al. (2001) X X X
Chan, Carter, and Burnes (2001) X X
2002 Wu et al. (2002) X X X
Angelelli and Speranza (2002) X X X
Zhang, Jiang, and Tang (2002) X X X
Giosa, Tansini, and Viera (2002) X X
2003 Dondo et al. (2003) X X X
Kazaz and Altinkemer (2003) X X X
2004 Matos and Oliveira (2004) X X
Wasner and Zapfel (2004) X X
Jin et al. (2004) X
2005 Mingozzi (2005) X X
Nagy and Salhi (2005) X X
Polacek et al. (2005) X X X
Lim and Wang (2005) X X X
Baltz, Dubhashi, Tansini, Srivastav, and Werth (2005) X
Songyan, Akio, and Bai (2005) X
Jin et al. (2005) X X
Songyan and Akio (2005) X
2006 Parthanadee and Logendran (2006) X X
Yang, Cui, and Cheng (2006) X X
Chiu et al. (2006) X X
Lim and Zhu (2006) X X X X
2007 Dondo and Cerd (2007) X X X X
Jeon et al. (2007) X X
Crevier et al. (2007) X X
Bae, Hwang, Cho, and Goan (2007) X X X
Pisinger and Ropke (2007) X X X X
 zyurt and Aksen (2007) X X
Carlsson, Ge, Subramaniam, Wu, and Ye (2007) X
Hu, Chen, and Guo (2007) X X
Wang, Gao, Cui, and Chen (2007) X
Lou (2007) X X X
Tsirimpas et al. (2007) X X X
2008 Ho et al. (2008) X X
Kek et al. (2008) X X X
Dondo et al. (2008) X X X X X
Polacek, Benkner, Doerner, and Hartl (2008) X X X
Dai, Chen, Pan, and Hu (2008) X X
Li and Liu (2008) X
Chen, Guo, and Fan (2008) X X
Yang (2008) X X X
Chen, Sheng-Zhang, and Shi (2008) X X X
Goela and Gruhn (2008) X X X X X X X
2009 Flisberg et al. (2009) X X X X X X

J.R. Montoya-Torres et al. / Computers & Industrial Engineering 79 (2015) 115129

Dondo and Cerd (2009) X X X X
Baldacci and Mingozzi (2009) X
Ting and Chen (2009) X X
Zhen and Zhang (2009) X X X
2010 Schmid et al. (2010) X X X X
Mirabi et al. (2010) X X
Liu et al. (2010) X X X
Vidal et al. (2010) X X X X
Ma and Yuan (2010) X
Sombunthama and Kachitvichyanukulb (2010) X X X
Sepehri and Kargari (2010) X X
Gajpal and Abad (2010) X X
Villegas et al. (2010) X X X X
Garaix, Artigues, Feillet, and Josselin (2010) X X X X
Ghoseiri and Ghannadpour (2010) X X X X
Kansou and Yassine (2010) X X
2011 Gulczynski et al. (2011) X X X
Bettinelli et al. (2011) X X X X X
Ycenur and Demirel (2011a) X
Aras et al. (2011) X X
Zarandi et al. (2011) X X X
Yang, Zhou, Cui, and He (2011) X X X
Wang et al. (2011) X X X X
Samanta and Jha (2011) X X X
Ycenur and Demirel (2011b) X X
Yu, Yang, and Xie (2011) X
Lei, Xing, Wu, and Wen (2011) X
Fard and Setak (2011) X X X
Zhang, Tang, and Fung (2011) X X
Surekha and Sumathi (2011) X X
Maischberger and Cordeau (2011) X X X X X X
2012 Cornillier et al. (2012) X X X X X
Kuo and Wang (2012) X X X
Lpez Franco and Nieto Isaza (2012) X X
Maya et al. (2012) X X X X
Nieto Isaza et al. (2012) X X X
Xu et al. (2012) X X X X
2013 Ben Alaia, Dridi, Bouchriha, and Borne (2013) X X X
Benavent and Martnez (2013) X
Liu (2013) X X X
Liu and Yu (2013) X X X X
Venkata et al. (2013) X X X

123
(continued on next page)
124 J.R. Montoya-Torres et al. / Computers & Industrial Engineering 79 (2015) 115129

employed by the enterprises associated with the research. The

Meta-heuristic

multi-objective MDVRP with time windows and split delivery is

studied by Dharmapriya and Siyambalapitiya (2010). The objec-
tives to be optimized were dened to be the total transportation
cost, the total distance traveled, full use of vehicle capacity and
X

X
X

X
X
X
X

X
X
load balancing. The problem is solved using Tabu Search, Simu-
lated Annealing and a Greedy algorithm. Tavakkoli-Moghaddam,
Heuristic
Solution method

Makui, and Mazloomi (2010) studied the multi-objective problem

in which depot location and routes of vehicles have to be dened
X
X

X
X
X
X
simultaneously. This problem is known in the literature as the
Exact

Multi-Depot Location-Routing Problem. Traditionally, this problem

X

X
X
X

X is solved sequentially: rst, the location of depots is addressed and

then the routing of vehicles is approached. These authors proposed
Other

a scatter search algorithm that seeks to maximize total demand

X

X
X
X
X
X
X

X
X
served and to minimize the total operational cost (cost of opening
depots and variable delivery costs). Computational experiments
Split delivery

showed that the proposed multi-objective scatter search (MOSS)

algorithm outperformed an Elite Tabu Search (ETS) procedure.
Scatter Search operates on a set of solutions, the reference set, by
X

combining these solutions to create new ones. Then, a subset is

Pick up & Delievery

non-randomly selected from this reference set and selected

solutions are combined to get starting solutions to run an improve-
ment mechanism. The reader interested in more details of the
multi-objective of Scatter Search can refer to the works of
Beausoleil (2001), Beausoleil (2004, Beausoleil (2006). On the other
hand the ETS algorithm considered by Tavakkoli-Moghaddam et al.
X
X

(2010) rst determines the minimum number of required facilities

Periodic

and then evaluates all combinations of facilities by incorporating

an elite tabu search procedure at the routing phase.
X

X
X

Jiang and Ding (2010) minimized the distribution cost, the

customer dissatisfaction and the changes of routes in a disruption
Capacitated

measurement model and an immune algorithm. The procedure is

tested using a simple example. Ghoseiri and Ghannadpour (2010)
considered the problem of simultaneously optimizing total eet
X
X
X
X

X
X

X
X

X
X

X
X

size and total distance deviation by using a genetic algorithm

Heterogeneous eet

coupled with goal programming. Finally, Venkatasubbaiah,

Acharyulu, and Chandra Mouli (2011) proposed the use of a Fuzzy
Goal Programming Method to solve the multi-objective problem. A
fuzzy max-min operator is also proposed to improve the effective-
ness of the procedure. The algorithm is tested on simple transpor-
X
X
X

X
X
X

tation problems from literature, and compared with previous

works, while Li and Liu (2011) proposed a genetic algorithm so
Time Windows

as to minimize the number of vehicles and the total travel distance.

Constraints

Adelzadeh et al. (in press) proposed a mathematical model and a

fuzzy-based heuristic for the particular case of the MDVRP with
fuzzy time windows and heterogeneous eet of vehicles. Mart-
X
X

X
X

nez-Salazar et al. (2014) studied the location-routing problem

(LRP) with transportation considerations in order to minimize
transport cost and load balancing. The bi-objective problem is for-
malized using mathematical programming and then solved with
Afshar-Nadja and Afshar-Nadja (in press)

both a Scatter Tabu Search procedure and the well-known NSGA-

II (Non-dominated Sorting Genetic Algorithm II). Mirabi (2014)
proposed a hybrid meta-heuristic algorithm coupling simulated
annealing and electromagnetism to minimize travel distance and
Contardo and Martinelli (2014)

customer waiting time for service. Finally, Rodrigues Pereira

Rahimi-Vahed et al. (2013)

Adelzadeh et al. (in press)

Subramanian et al. (2012)

Ramos, Gomes, and Barbosa-Pvoa (2014) simultaneously evalu-

Escobar et al. (in press)

Vahed et al. (in press)

Contardo et al. (2014)

ated travel distance and level of carbon emissions in a multi-depot

Braekers et al. (2014)

Luo and Chen (2014)

Xu and Jiang (2014)

Ray et al. (in press)
Hirsch et al. (2014)

vehicle routing problem found in waste collection supply chain.

Vidal et al. (2014)
Li, et al. (in press)

Sitek et al. (2014)

Salhi et al. (2014)
Seth et al. (2013)

Wang (2013)
Reference
Table A1 (continued)

4. Analysis of literature

As pointed out before, since the rst publication on multi-depot

vehicle routing problem appeared in 1984. Since there, more than
2014
Year

145 papers have been published up to date in the scientic

literature about this problem and its variants. Between the middle
 J.R. Montoya-Torres et al. / Computers & Industrial Engineering 79 (2015) 115129 125

of the 1980s (when the rst works on the MDVRP were published)
Meta-heuristic and the end of the twentieth century, very few papers were pro-
posed in the literature (see Fig. 3): only 18 publications (excluding
those published in 2000), giving an average of 1.12 papers per year.
82%

Between 2000 and 2005, there was an increase in the number of

X
X
X
X
X
X
X

X
X
X

X
X
X
X
publication on MDVRP with an average of 4.33 publications per
year. This gives a total of 44 publications until 2005 inclusive.
Heuristic

The most impressive growing on the number of papers published

Solution method

35%

is observed between 2006 and 2014 with a total 103 publications,

X

X
X
X

X
giving an average of 10.4 papers per year in the period 20062010
and 12.75 paper per year in the period 20112014 inclusive.
Exact
12%

As can be seen, there is a clearly increasing trend showing the

X

X
growing interest in this eld. It is reasonable to expect that in
the coming years the MDVRP will receive an ever large amount
Other
41%

of attention. There are however some remarks to be made. As

X

X
X

X
shown in Fig. 4, most of the works have been focused on the min-
imization of cost, distance or time. The papers dealing with vehicle
Split delivery

load balancing are in fact papers that seek to optimize multiple

objectives simultaneously (very often cost and vehicle load). As
presented in previous sections of this review (see also Fig. 4), the
6%

majority of published works deals with the single objective

problem. While this problem is of theoretical interest, very often
Pick up & Delievery

decision-makers are faced to optimize multiple (contradictory)

objectives. Very few published works deals with multi-objective
problem. Hence, this gap in current research could be closed by
proposing efcient and effective solution approaches for multi-
objective environments.
12%

It is also interesting to study the different methodologies and

techniques that the authors apply in the reviewed literature.
Periodic

Fig. 5 shows two pie charts with this distribution: it rst classies
12%

the approaches as exact, heuristics and meta-heuristics algorithms,

X

while the second pie presents a distribution of the different

approximate algorithms employed in the reviewed literature. We
Capacitated

rst observe that exact algorithms (branch and bound, mathemat-

ical programming) are employed in 25% of reviewed papers. We
41%

have to consider that these techniques have proven to be useful

X

X
X
X
X

for simplied combinatorial optimization problems, specic

settings and/or small instances. Hence, a larger focus is needed
Heterogeneous eet

on approaches able to solve larger instances. Because of the NP-

hardness of the MDVRP, approximate heuristics have also been
proposed. Under the category of heuristic algorithms, we have
classied many different algorithms and ad-hoc methods that are
12%

specic and do not contain a well-known meta-heuristic template.

X

This represents 33% of the reviewed papers. For the other 42% a
meta-heuristic algorithm is proposed. Among the available proce-
Time windows

dures, Tabu Search (TS), genetic algorithms (GA) and simulated

Constraints

annealing (SA) have been the most employed in the reviewed

literature. The other meta-heuristics have been less employed:
29%

ant colony optimization and variable neighborhood search. Clearly,

X

X
X

there is a large opportunity for research here. Meta-heuristics have

Dharmapriya and Siyambalapitiya (2010)

long ago established themselves as state-of-the-art methodologies

Reviewed papers about the multi-objective MDVRP.

Rodrigues Pereira Ramos et al. (2014)

for the vast majority of vehicle routing problems.

Ombuki-Berman and Hanshar (2009)

Tavakkoli-Moghaddam et al. (2010)

Ghoseiri and Ghannadpour (2010)

Venkatasubbaiah et al. (2011)

Martnez-Salazar et al. (2014)

5. Conclusions and lines for future research

Adelzadeh et al. (in press)
Hasanpour et al. (2009)

Jiang and Ding (2010)

Lin and Kwok (2005)

Finding an optimal (or at least very good) route for vehicles

Weise et al. (2010)

Li and Liu (2011)

delivering products to customers is one of the key functions in

Tan et al. (2006)
Lau et al. (2009)

Lau et al. (2010)

Mirabi (2014)

any logistics systems. Since the rst works proposed in literature

% of papers

in the late 1959, the literature on vehicle routing problems has

Reference

been highly increasing. A lot of works, as well as state-of-the-art

surveys, have been published. In this paper, we focused on one of
the less reviewed variants: the vehicle routing problem with multi-
ple depots (MDVRP). We presented a complete analysis of scientic
Table A2

2011

2014
2005
2006
2009

2010
Year

literature since the rst works about this problem were published
in the mid 1980s.
126 J.R. Montoya-Torres et al. / Computers & Industrial Engineering 79 (2015) 115129

Early works on MDVRP mainly proposed exact algorithms References

(mathematical models, branch-and-bound methods). With the
development of computers and the increasing popularity of Adelzadeh, M., Asi, V. M., & Koosha, M. (in press). A mathematical model and solving
procedure for the multi-depot vehicle routing problem with fuzzy time window
meta-heuristics, researches then focused on proposing genetic and heterogeneous vehicle. International Journal of Advanced Manufacturing
algorithms, tabu search procedures, simulated annealing, ant Technology. doi: http://dx.doi.org/10.1007/s00170-014-6141-8.
colony optimization algorithms or even variable neighborhood Afshar-Nadja, B., & Afshar-Nadja, A. (2014). A constructive heuristic for time-
dependent multi-depot vehicle routing problem with time-windows and
procedures to efciently solve the problem. Some hybrid proce- heterogeneous eet. Journal of King Saud University Engineering Sciences. doi:
dures have been proposed and their effectiveness has been proved http://dx.doi.org/10.1016/j.jksues.2014.04.007.
by using benchmark instances. It is to note that several meta-heu- Angelelli, E., & Speranza, M. G. (2002). The periodic vehicle routing problem with
intermediate facilities. European Journal of Operational Research, 137(2),
ristics (like GRASP), matheuristics, simheuristics and other hybrid
233247.
procedures can be exploited to solve the problem. Also, as complex Aras, N., Aksen, D., & Tekin, M. T. (2011). Selective multi-depot vehicle routing
realistic problems appear to be of interest for researchers, addi- problem with pricing. Transportation Research Part C, 19(5), 866884.
tional constraints (product pickup and delivery, vehicle capacity Archetti, C., & Speranza, M. G. (2008). The split delivery vehicle routing problem: a
survey. In The vehicle routing problem: latest advances and new challenges.
constraints, delivery time windows, etc.) have been added to the Operations research/computer science interfaces Series: Vol. 43. Part I (pp. 103
original problem. 122).
Most of the rst works on MDVRP looked at the minimization of Bae, S.-T., Hwang, H. S., Cho, G.-S., & Goan, M.-J. (2007). Integrated GA-VRP solver for
multi-depot system. Computers & Industrial Engineering, 53(2), 233240.
travel cost, distance or time. Very few papers considered nontradi- Baldacci, R., Battarra, M., & Vigo, D. (2008). Routing a heterogeneous eet of
tional objective functions. There is an interesting gap in literature vehicles. In B. L. Golden, S. Raghavan, & E. Wasil (Eds.), The vehicle routing
here since many other objective functions are of industrial rele- problem: latest advances and new challenges (pp. 327). Berlin: Springer. Vol. 43.
Baldacci, R., Toth, P., & Vigo, D. (2007). Recent advances in vehicle routing exact
vance. In addition, logistics problems are multi-objective (MO) by algorithms. 4OR, 5(4), 269298.
nature, which means that several criteria, very often in conict Baldacci, R., & Mingozzi, A. (2009). A unied exact method for solving different
with each other, have to be considered at a time. Very few pub- classes of vehicle routing problems. Mathematical Programming, 120, 347380.
Baldacci, R., Toth, P., & Vigo, D. (2010). Exact algorithms for routing problems under
lished works on multi-depot vehicle routing problems deals with vehicle capacity constraints. Annals of Operations Research, 175, 213245.
multiple objectives. Research in multi-objective optimization is Baltz, A., Dubhashi, D., Tansini, L., Srivastav, A., & Werth, S. (2005). Probabilistic
concerned with the generation of solutions in which none of the analysis for a multiple depot vehicle routing problem. Foundations of Software
Technology and Theoretical Computer Science, Lecture Notes in Computer Science,
objective functions can be improved without paying a cost in other
3821(2005), 360371.
objective(s) (usually referred to as non-dominated solutions). Beausoleil, R. P. (2004). Bounded variables nonlinear multiple criteria optimization
Hence, researches must focus on proposing efcient and effective using scatter search. Revista de Matemtica: Teora y Aplicaciones, 11(1), 1740.
decision-making tools for multi-objective environments. The deal Beausoleil, R. P. (2006). MOSS multiobjective scatter search applied to non-linear
multiple criteria optimization. European Journal of Operational Research, 169(2),
is not to consider several objectives and to optimize them sequen- 426449.
tially (e.g., by concentrating on optimizing one objective and then, Beausoleil, R. P. (2001). Multiple criteria scatter search. In J. Pinho de Sousa (Ed.),
with that solution, considering the other objectives as secondary. Proceedings of 4th Metaheuristics International Conference (pp. 534539).
Portugal: Porto.
The attempt is to nd non-dominated solutions. Ben Alaia, E., Dridi, I. H., Bouchriha, H., & Borne, P. (2013). Optimization of the multi-
Finally, with the increasingly worldwide concern on the optimal depot & Multi-vehicle pickup and delivery problem with time windows using
use of scarcer natural resources, enterprises and supply chains genetic algorithm. In Proceedings of the 2013 International Conference on Control,
Decision and Information Technologies (CoDIT 2013) (pp. 343348).
need to be re-aligned to adjust to this trend. In order to go further Benavent, E., & Martnez, A. (2013). Multi-depot multiple TSP: A polyhedral study
from the classical economic optimization, the inclusion of social and computational results. Annals of Operations Research, 207(1), 725.
and environmental metrics in optimization problems becomes Bettinelli, A., Ceselli, A., & Righini, G. (2011). A branch-and-cut-and-price algorithm
for the multi-depot heterogeneous vehicle routing problem with time windows.
highly relevant. The paper of Montoya-Torres (2014) presents a Transportation Research Part C, 19(5), 723740.
general framework to take into account sustainability issues in Bodin, L. (1975). A taxonomic structure for vehicle routing and scheduling
supply chain decision-making, including logistics distribution problems. Computers and Urban Society, 1, 1129.
Bodin, L., & Golden, B. (1981). Classication in vehicle routing and scheduling.
process (i.e. vehicle routing decisions), while the work of
Networks, 11, 97108.
Montoya-Torres, Muoz-Villamizar, and Vega-Meja (2014) seeks Braekers, K., Caris, A., & Jenssens, G. K. (2014). Exact and meta-heuristic approach
at proposing the application of the MDVRP in city logistics for for a general heterogeneous dial-a-ride problem with multiple depots.
the minimization of travel costs and carbon emissions. Some rst Transportation Research Part B: Methodological, 67, 166186.
Carlsson, J., Ge, D., Subramaniam, A., Wu, A., & Ye, Y. (2007). Solving Min-Max Multi-
attempts to consider sustainable issues (i.e. economic, environ- Depot Vehicle Routing Problem. In Proceedings of the 2007 Fields Institute
mental and social issues simultaneously) when solving logistic Workshop on Global Optimization. Available online at: <http://
distribution problems have been presented in literature, see e.g. www.stanford.edu/~yyye/MDVRP-JGSWY.pdf>. Accessed 1 May 2012.
Carpaneto, G., Dellamico, M., Fischetti, M., & Toth, P. (1989). A branch and bound
(Sbihi & Eglese, 2007). This certainly leads to solve more complex algorithm for the multiple depot vehicle scheduling problem. Networks, 19(5),
multi-objective routing problems. 531548.
Chan, Y., Carter, W. B., & Burnes, M. D. (2001). A multiple-depot, multiple-vehicle,
location-routing problem with stochastically processed demands. Computers &
Acknowledgments Operations Research, 28(8), 803826.
Chao, I.-M., Golden, B. L., & Wasil, E. (1993). A new heuristic for the multi-depot
vehicle routing problem that improves upon best-known solutions. American
Authors would like to thanks the anonymous referees for their
Journal of Mathematical and Management Sciences, 13(34), 371406.
suggestions allowing the improvement of this manuscript. Part of Chen, X., Guo, Q., & Fan, C. (2008). Optimal algorithm for multiple-depot vehicle
this work was supported by the Colombian Department of Science, routing problem with full loads. In Computer Engineering and Design, 2008-22.
Technology and Innovation (Colciencias) under Grant No. 333-502- Available online at: <http://en.cnki.com.cn/Article_en/CJFDTOTAL-
SJSJ200822062.htm>. Accessed 1 May 2012.
27900, and by Universidad de La Sabana, Colombia, under Research Chen, M.-J., Sheng-Zhang, Z., & Shi, J.-S. (2008). Study on ant colony algorithm for
Project Modeling and analysis of production and transport multi-depots vehicle routing problem with multiple constraints. Zhongguo Jixie
systems with economic, social and environmental metrics. Gongcheng/China Mechanical Engineering, 19(16), 19391944.
Chen, K., Takama, Y., & Hirota, K. (2000). A Hierarchical Multiplex Structure Model
for Practical Transportation Problems with Multiple Depots. International
Journal of Fuzzy Systems, 2, 305314.
Appendix A.
Chiu, H. N., Lee, Y. S., & Chang, J. H. (2006). Two approaches to solving the multi-
depot vehicle routing problem with time windows in a time-based logistics
(see Tables A1 and A2). environment. Production Planning & Control, 17(5), 480493.
 J.R. Montoya-Torres et al. / Computers & Industrial Engineering 79 (2015) 115129 127

Christodes, N., & Eilon, S. (1969). An algorithm for the vehicle-dispatching categorized bibliography. In B. Golden, S. Raghavan, & E. Wasil (Eds.), The
problem. Journal of the Operational Research Society, 20, 309318. vehicle routing problem: latest advances and new challenges. Operations Research/
Clarke, G., & Wright, J. W. (1964). Scheduling of vehicles from a depot to a number Computer Science Interfaces Series, Part I (Vol. 43, pp. 143169). New York:
of delivery points. Operations Research, 12, 568581. Springer.
Contardo, C., Cordeau, J.-F., & Gendron, B. (2014). An exact algorithm based on cut- Ghoseiri, K., & Ghannadpour, S. F. (2010). Multi-objective vehicle routing problem
and-column generation for the capacitated location-routing problem. INFORMS with time windows using goal programming and genetic algorithm. Applied Soft
Journal on Computing, 26(1), 88102. Computing, 10, 10961107.
Contardo, C., & Martinelli, R. (2014). A new exact algorithm for the multi-depot Gillett, B., & Johnson, J. (1976). Multi-terminal vehicle-dispatch algorithm. Omega,
vehicle routing problem under capacity and route length constraints. Discrete 4(6), 711718.
Optimization, 12, 129146. Giosa, I. D., Tansini, I. L., & Viera, I. O. (2002). New assignment algorithms for the
Cordeau, J. F., Laporte, G., Savelsbergh, M. W. P., & Vigo, D. (2007). Vehicle routing. In multi-depot vehicle routing problem. Journal of the Operational Research Society,
C. Barnhart & G. Laporte (Eds.). Transportation, handbooks in operations research 53, 977984.
and management science (Vol. 14, pp. 367428). Amsterdam: North-Holland. Goela, A., & Gruhn, V. (2008). A general vehicle routing problem. European Journal of
Cordeau, J. F., Gendreau, M., Hertz, A., Laporte, G., & Sormany, J. S. (2005). New Operational Research, 191(3), 650660.
heuristics for the vehicle routing problem. In A. Langevin & D. Riopel (Eds.), Gulczynski, D., Golden, B. L., & Wasil, E. (2011). The multi-depot split delivery
Logistics Systems: Design and Optimization (pp. 279298). New York: Springer. vehicle routing problem: An integer programming-based heuristic, new test
Cordeau, J. F., Gendreau, M., & Laporte, G. (1997). A tabu search heuristic for problems, and computational results. Computers & Industrial Engineering, 61(3),
periodic and multi-depot vehicle routing problems. Networks, 30(2), 105119. 794804.
Cordeau, J. F., Laporte, G., & Mercier, A. (2001). A Unied Tabu Search Heuristic For Hadjiconstantinou, E., & Baldacci, R. (1998). A multi-depot period vehicle routing
Vehicle Routing Problems With Time Windows. Journal of the Operational problem arising in the utilities sector. Journal of the Operational Research Society,
Research Society, 52(8), 928936. 49(12), 12391248.
Cornillier, F., Boctor, F., & Renaud, J. (2012). Heuristics for the multi-depot petrol Hasanpour, H. A., Mosadegh-Khah, M., & Tavakoli Moghadam, R. (2009). Solving a
station replenishment problem with time windows. European Journal of stochastic multi-depot multi-objective vehicle routing problem by a simulated
Operational Research, 220, 361369. annealing. Journal of Industrial Engineering, 43(1), 2536.
Crevier, B., Cordeau, J. F., & Laporte, G. (2007). The multi-depot vehicle routing Hirsch, M. J., Schroeder, D. E., Maggiar, A., & Dolinskaya, I. S. (2014). Multi-depot
problem with inter-depot routes. European Journal of Operational Research, vessel routing problem in a direction dependent waveeld. Journal of
176(2), 756773. Combinatorial Optimization, 28, 3857.
Dai, S., Chen, W., Pan, Y., & Hu, Y. (2008). A hybrid ant colony algorithm for multiple Ho, W., Ho, G. T. S., Ji, P., & Lau, H. C. W. (2008). A hybrid genetic algorithm for the
depot vehicle routing problem. Journal of Sichuan University (Engineering Science multi-depot vehicle routing problem. Engineering Applications of Articial
Edition), 2008-06. Available online at: <http://en.cnki.com.cn/Article_en/ Intelligence, 21(4), 548557.
CJFDTOTAL-SCLH200806027.htm>. Accessed 1 May 2012. Hu, D., Chen, C., & Guo, X. (2007). Research on optimal algorithm for multi-depot
Dantzig, G., & Ramser, J. (1959). The truck dispatching problem. Management vehicle routing problem with pickups and deliveries. Mathematics in practice
Science, 6, 8091. and theory, 2007-02. Available at: <http://en.cnki.com.cn/Article_en/
Desrochers, M., Lenstra, J. K., & Savelsbergh, M. W. P. (1990). A classication scheme CJFDTOTAL-SSJS200702016.htm>. Accessed 1 May 2012.
for vehicle routing and scheduling problems. European Journal of Operational Irnich, S. (2000). A multi-depot pickup and delivery problem with a single hub and
Research, 46, 322332. heterogeneous vehicles. European Journal of Operational Research, 122(2), 310328.
Dharmapriya, U. S. S., & Siyambalapitiya, S. B. (2010). Articial intelligence Jeon, G., Leep, H. R., & Shim, J. Y. (2007). A vehicle routing problem solved by using
computational techniques to optimize a multi objective oriented distribution hybrid genetic algorithm. Computers & Industrial Engineering, 53(4), 680692.
operations. In Proceedings of the 2010 international conference on industrial Jiang, L., & Ding, B. (2010) Disruption management model of multi-depot vehicle
engineering and operations management CD-ROM. Dhaka, Bangladesh. Available routing problem with customers demand changes. Systems Engineering, 2010-
online at: <http://www.iieom.org/paper/195%20Subodha.pdf>. Accessed: 11 12. Available online at: <http://en.cnki.com.cn/Article_en/CJFDTOTAL-
November 2011. GCXT201012002.htm>. Accessed 30 May 2012.
Dondo, R., & Cerd, J. (2007). A cluster-based optimization approach for the multi- Jin, T., Guo, S., Wang, F., & Lim, A. (2004). One-stage search for multi-depot vehicle
depot heterogeneous eet vehicle routing problem with time windows. routing problem. Proceedings of the Information Systems and Control Conference.
European Journal of Operational Research, 176(3), 14781507. 446-129.
Dondo, R., & Cerd, J. (2009). A hybrid local improvement algorithm for large scale Jin, O. S., Mitsuo, G., & Hiroshi, K. (2005). Multi-depot vehicle routing problem with
multi-depot vehicle routing problems with time windows. Computers & time windows by hybrid genetic algorithm. Nippon Keiei Kogakkai Shunki Taikai
Chemical Engineering, 33, 513530. Yokoshu, 2005, 158159.
Dondo, R., Mendez, C. A., & Cerd, J. (2003). An optimal approach to the multiple- Juan, A. A., Fauln, J., Adelanteado, F., Grasman, S., & Montoya Torres, J. R. (2009).
depot heterogeneous vehicle routing problem with time window and capacity Solving the capacitated vehicle routing problem maximum traveling distance
constraints. Latin American Applied Research, 33(2), 129134. and service time requirements: An approach based on Monte Carlo. In M. D.
Dondo, R., Mndez, C. A., & Cerd, J. (2008). Optimal management of logistic Rossetti, R. R. Hill, B. Johanson, A. Dunkin, & R. G. Ingalls (Eds.) Proceedings of the
activities in multi-site environments. Computers & Chemical Engineering, 32(11), 2009 winter simulation conference (pp. 24672475).
25472569. Kansou, A., & Yassine, A. (2010). New upper bounds for the multi-depot capacitated
Eksioglu, B., Volkan, Vural. A., & Reisman, A. (2009). The vehicle routing problem: A arc routing problem. International Journal of Metaheuristics, 1(1), 8195.
taxonomic review. Computers & Industrial Engineering, 57, 14721483. Kazaz, B., & Altinkemer, K. (2003). Optimization of multi-feeder (depot) printed
Escobar, J. W., Linfati, R. Toth, P., & Baldoquin, M. G. (in press). A hybrid granular circuit board manufacturing with error guarantees. European Journal of
tabu search algorithm for the multi-depot vehicle routing problem. Journal of Operational Research, 150(2), 370394.
Heuristics. doi: http://dx.doi.org/10.1007/s10732-014-9247-0. Kek, A., Cheu, R., & Meng, Q. (2008). Distance-constrained capacitated vehicle
Fard, F. A., & Setak, M. (2011). Comparison between two algorithms for multi-depot routing problems with exible assignment of start and end depots.
vehicle routing problem with inventory transfer between depots in a three- Mathematical and Computer Modelling, 47(12), 140152.
echelon supply chain. International Journal of Computer Applications, 28(6), Kulkarni, R. V., & Bhave, P. R. (1985). Integer programming formulations of vehicle
3945. routing problems. European Journal of Operational Research, 20(1), 5867.
Filipec, M., Skrlec, D., & Krajcar, S. (1997). Darwin meets computers: new approach Kuo, Y., & Wang, C.-C. (2012). A variable neighborhood search for the multi-depot
to multiple depot capacitated vehicle routing problem. In Proceedings of the vehicle routing problem with loading cost. Expert Systems with Applications,
1997 IEEE international conference on systems, man, and cybernetics, Vol. 1 (pp. 39(8), 69496954.
421426). Laporte, G. (1989). Integer programming formulations for the multi-depot vehicle
Filipec, M., Skrlec, D., & Krajcar, S. (2000). Genetic algorithm approach for multiple routing problem: Comments on a paper by Kulkarni and Bhave. European
depot capacitated vehicle routing problem solving with heuristics Journal of Operational Research, 38(2), 228237.
improvements. International Journal of Modeling and Simulation, 20(4), 320328. Laporte, G. (1992). The vehicle routing problem: An overview of exact and
Flisberg, P., Lidn, B., & Rnnqvist, M. (2009). A hybrid method based on linear approximate algorithms. European Journal of Operational Research, 59, 345358.
programming and tabu search for routing of logging trucks. Computers & Laporte, G., & Semet, F. (2002). Classical heuristics for the capacitated VRP. In P.
Operations Research, 36(4), 11221144. Toth, & D. Vigo (Eds.), The vehicle routing problem. SIAM monographs on discrete
Gajpal, Y., & Abad, P. L. (2010). Saving based algorithm for multi-depot version of mathematics and applications, Vol. 9 (pp. 109128). Philadelphia: SIAM.
vehicle routing problem with simultaneous pickup and delivery. International Laporte, G., & Nobert, Y. (1987). Exact algorithms for the vehicle routing problem.
Journal of Enterprise Network Management, 3(3), 201222. Annals of Discrete Mathematics, 31, 147184.
Garaix, T., Artigues, C., Feillet, D., & Josselin, D. (2010). Vehicle routing problems Laporte, G., Nobert, Y., & Arpin, D. (1894). Optimal solutions to capacitated
with alternative paths: An application to on-demand transportation. European multidepot vehicle routing problems. Congressus Numerantium, 44, 283292.
Journal of Operational Research, 204(1), 6275. Laporte, G., Nobert, Y., & Taillefer, S. (1988). Solving a family of multi-depot vehicle
Garey, M. R., & Johnson, D. S. (1979). Computers and intractability: A guide to the routing and location-routing problems. Transportation Science, 22(3), 161172.
theory of NP-completeness. New York, USA: W.H. Freeman & Co Ltd. Lau, H. C. W., Chan, T. M., Tsui, W. T., & Pang, W. K. (2010). Application of genetic
Gendreau, M., Laporte, G., & Potvin, J.-Y. (2002). Metaheuristics for the capacitated algorithms to solve the multidepot vehicle routing problem. IEEE Transactions
VRP. In P. Toth & D. Vigo (Eds.), The vehicle routing problem. SIAM monographs on on Automation Science and Engineering, 7(2), 383392.
discrete mathematics and applications (Vol. 9, pp. 129154). Philadelphia: SIAM. Lau, H. C. W., Tang, C. X. H., Ho, G. T. S., & Chan, T. M. (2009). A fuzzy guided multi-
Gendreau, M., Potvin, J.-Y., Brumlaysy, O., Hasle, G., & Lkketangen, A. (2008). objective evolutionary algorithm model for solving transportation problem.
Meta-heuristics for the vehicle routing problem and its extensions: A Expert Systems with Applications, 36(4), 82558268.
128 J.R. Montoya-Torres et al. / Computers & Industrial Engineering 79 (2015) 115129

Lei, X. M., Xing, C. F., Wu, L., & Wen, Y. F. (2011). Multiple-depot vehicle routing Montoya-Torres, J. R., Muoz-Villamizar, S., & Vega-Meja, C. A. (2014). On the
problems as a distributed constraint optimization problem. Applied Mechanics impact of collaborative strategies for goods delivery in city logistics. Working
and Materials, 6668, 10331038. Paper, Universidad de La Sabana, Chia, Colombia. Submitted to Production
Li, J., Li, Y., & Pardalos, P. M. (in press). Multi-depot vehicle routing problem with Planning & Control.
time windows under shared depot resources. Journal of Combinatorial Mourgaya, M., & Vanderbeck, F. (2006). The periodic vehicle routing problem:
Optimization. doi: http://dx.doi.org/10.1007/s10878-014-9767-4. Classication and heuristic for tactical planning. RAIRO Operations Research,
Li, Y., & Liu, X. (2008). Modelling and its ant colony algorithm for multi-depot open 40, 169194.
vehicle routing problem with replenishment on the way. Computer Integrated Nagy, G., & Salhi, S. (2005). Heuristic algorithms for single and multiple depot
Manufacturing Systems, 2008-03. Available online at: <http://en.cnki.com.cn/ vehicle routing problems with pickups and deliveries. European Journal of
Article_en/CJFDTOTAL-JSJJ200803022.htm>. Accessed: November 11, 2011. Operational Research, 162(1), 126141.
Li, Y., & Liu, C. (2011). An effective genetic algorithm for the vehicle routing problem Narasimha, K. V., Kivelevitch, E., Sharma, B., & Kumar, M. (2013). An ant colony
with multiple depots. Advanced Materials Research, 204210, 283287. optimization technique for solving min-max multi-depot vehicle routing
Lim, A., & Zhu, W. (2006). A fast and effective insertion algorithm for multi-depot problem. Swarm and Evolutionary Computation, 13, 6373.
vehicle routing problem and a new simulated annealing approach. In M. Ali & R. Nieto Isaza, S., Lpez Franco, J., & Herazo Padilla, N. (2012). Desarrollo y Codicacin
Dapoigny (Eds.), Advances in articial intelligence. Lecture notes in computer de un Modelo Matemtico para la Optimizacin de un Problema de Ruteo de
science (Vol. 4031, pp. 282291). Springer-Verlag. Vehculos con Mltiples Depsitos. In Tenth LACCEI Latin American and
Lim, A., & Wang, F. (2005). Multi-depot vehicle routing problem: A one-stage Caribbean Conference (LACCEI2012), Megaprojects: Building Infrastructure by
approach. IEEE Transactions on Automation Science and Engineering, 2(4), 397402. fostering engineering collaboration, efcient and effective integration and
Lin, C. K. Y., & Kwok, R. C. W. (2005). Multi-objective metaheuristics for a location innovative planning , July 2327, 2012, Panama City, Panama. Accepted.
routing problem with multiple use of vehicles on real data and simulated data. Ombuki-Berman, B., & Hanshar, F. T. (2009). Using genetic algorithms for multi-
European Journal of Operational Research, 175(3), 18331849. depot vehicle routing. Bio-inspired Algorithms for the Vehicle Routing Problem
Liong, C. Y., Wan, R. I., Khairuddin, O., & Mourad, Z. (2008). Vehicle routing problem: Studies in Computational Intelligence, 161(2009), 7799.
Models and solutions. Journal of Quality Measurement and Analysis, 4(1), Ozrat, P. M., & Ozkarahan, I. (2010). A constraint programming heuristic for a
205218. heterogeneous vehicle routing problem with split deliveries. Applied Articial
Liu, C.-Y. (2013). An improved adaptive genetic algorithm for the multi-depot Intelligence, 24(4), 277294.
vehicle routing problem with time windows. Journal of Networks, 8(5), zyurt, Z., & Aksen, D. (2007). Solving the multi-depot location routing problem
10351042. with Lagrangian relaxation. Operations Research/Computer Science Interfaces
Liu, T., Jiang, Z., Liu, R., & Liu, S. (2011). A review of the multi-depot vehicle routing Series, 37, 125144.
problem (Proc. of ESEP 2011 Conference). Energy Procedia, 13, 33813389 Parthanadee, P., & Logendran, R. (2006). Periodic product distribution from multi-
(Singapore, 910 December 2011). depots under limited supplies. IIE Transactions, 38(11), 10091026.
Liu, R., Jiang, Z., Fung, R. Y. K., Chen, F., & Liu, X. (2010). Two-phase heuristic Pisinger, D., & Ropke, S. (2007). A general heuristic for vehicle routing problems.
algorithms for full truckloads multi depot capacitated vehicle routing problem Computers & Operations Research, 34(8), 24032435.
in carrier collaboration. Computers & Operations Research, 37(5), 950959. Polacek, M., Benkner, S., Doerner, K. F., & Hartl, R. F. (2008). A cooperative and
Liu, C., & Yu, J. (2013). Multiple depots vehicle routing based on the ant colony with adaptive variable neighborhood search for the multi depot vehicle routing
the genetic algorithm. Journal of Industrial Engineering and Management, 6(4), problem with time windows. Business Research, 1(2), 207218.
10131026. Polacek, M., Hartl, R. F., Doerner, K., & Reimann, M. (2005). A variable neighborhood
Lpez Franco, J., & Nieto Isaza, S. (2012). Heurstica para la generacin de un search for the multi depot vehicle routing problem with timewindows. Journal
conjunto de referencia de soluciones que resuelvan el problema de ruteo de of Heuristics, 10(6), 613627.
vehculos con mltiples depsitos MDVRP. In Tenth LACCEI Latin American and Psaraftis, H. (1995). Dynamic vehicle routing: Status and prospects. Annals of
Caribbean Conference (LACCEI2012), Megaprojects: Building Infrastructure by Operations Research, 61(1), 143164.
fostering engineering collaboration, efcient and effective integration and Rahimi-Vahed, A., Crainic, T. G., Gendreau, M., & Rei, W. (2013). A path relinking
innovative planning , July 2327, 2012, Panama City, Panama. Accepted.. algorithm for a multi-depot periodic vehicle routing problem. Journal of
Lpez-Castro, L. F., & Montoya-Torres, J. R. (2011). Vehicle routing problem with Heuristics, 19, 497524.
fuzzy time windows. In Proceedings of the 2011 IEEE workshop on computational Ray, S., Soeanu, A., Berger, J., & Debbabi, M. (in press). The multi-depot split-delivery
intelligence in production and logistics systems (pp. 3946). IEEE Publishing. vehicle routing problem: Model and solution algorithm. Knowledge-based
Lou, S. (2007). Method for solving multi-depot stochastic vehicle routing problems. Systems. doi: http://dx.doi.org/10.1016/j.knosys.2014.08.006.
Journal of System Simulation, 2007-04. Available online at: <http:// Renaud, J., Boctor, F. F., & Laporte, G. (1996). An improved petal heuristic for the
en.cnki.com.cn/Article_en/CJFDTOTAL-XTFZ200704044.htm>. Accessed 1 May vehicle routing problem. Journal of the Operational Research Society, 47, 329336.
2012. Renaud, J., Laporte, G., & Boctor, F. F. (1996). A Tabu search heuristic for the multi-
Luo, J., & Chen, M.-R. (2014). Improved shufed frog leaping algorithm and its multi- depot vehicle routing problem. Computers & Operations Research, 23(3),
phase model for multi-depot vehicle routing problem. Expert Systems with 229235.
Applications, 41, 25352545. Rodrigues Pereira Ramos, T., Gomes, M. I., & Barbosa-Pvoa, A. P. (2014). Economic
Ma, J., & Yuan, J. (2010). Ant colony algorithm for multiple-depot vehicle routing and environmental concerns in planning recycle waste collection systems.
problem with shortest nish time. E-business Technology and Strategy: Transportation Research Part E: Logistics and Transportation Review, 62, 3454.
Communications in Computer and Information Science, 113, 114123. Salhi, S., Imran, A., & Wassan, N. A. (2014). The multi-depot vehicle routing problem
Mafoli, F. (2002). The vehicle routing problem: A book review. 4OR: A Quarterly with heterogeneous vehicle eet: Formulation and a variable neighborhood
Journal of Operations Research, 1(2), 149153. search implementation. Computers & Operations Research, 52, 315325.
Maischberger M., & Cordeau J. -F. (2011). Solving variants of the vehicle routing Salhi, S., & Nagy, G. (1999). A cluster insertion heuristic for single and multiple
problem with a simple parallel iterated Tabu Search. In Network optimization. depot vehicle routing problems with backhauling. Journal of the Operational
Lecture notes in computer science, Vol. 6701 (pp. 395-400). Research Society, 50(10), 10341042.
Matos, A. C., & Oliveira, R. C. (2004). An experimental study of the ant colony system Salhi, S., & Sari, M. (1997). A multi-level composite heuristic for the multi-depot
for the period vehicle routing problem. Ant colony optimization and swarm vehicle eet mix problem. European Journal of Operational Research, 103(1),
intelligence. In M. Dorigo, M. Birattari, C. Blum, L.M. Gambardella, F. Mondada, 95112.
& T. Sttzle (Eds.), Lecture Notes in Computer Science, Vol. 3172 (pp. 129). Salhi, S., Thangiah, S. R., & Rahman, F. (1998). A genetic clustering method of the
Maya, P., Srensen, K., & Goos, P. (2012). A metaheuristic for a teaching assistant multi-depot vehicle routing problem. In G. D. Smith & N. C. Steel (Eds.), Articial
assignment-routing problem. Computers & Operations Research, 39(2), 249258. neural network and genetic algorithms (pp. 234237). NY: Springer.
Min, H., Current, J., & Schilling, D. (1992). The multiple depot vehicle routing Samanta, S., & Jha, M. K. (2011). Multi depot probabilistic vehicle routing problems
problem with backhauling. Journal of Business Logistics, 13(1), 259288. with a time window: Theory, solution and application. International Journal of
Mingozzi, A. (2005). The multi-depot periodic vehicle routing problem. In J. -D. Operations Research and Information Systems, 2(2), 4064.
Zucker, & L. Saitta (Eds.), Abstraction, reformulation and approximation. Lecture Sbihi, A., & Eglese, R. W. (2007). Combinatorial optimization and Green Logistics.
notes in computer science, Vol. 3607 (pp. 347350). Berlin Heidelberg: Springer. 4OR: A Quarterly Journal of Operations Research, 5(2), 99116.
Mirabi, M. (2014). A hybrid electromagnetism algorithm for multi-depot periodic Schmid, V., Doerner, K. F., Hartl, R. F., & Salazar-Gonzlez, J. J. (2010). Hybridization
vehicle routing problem. International Journal of Advanced Manufacturing of very large neighborhood search for ready-mixed concrete delivery problems.
Technology, 71, 509518. Computers & Operations Research, 37(3), 559574.
Mirabi, M., Fatemi Ghomi, S. M. T., & Jolai, F. (2010). Efcient stochastic hybrid S
en, A., & Blbl, K. (2008). A survey on multi trip vehicle routing problem. In
heuristics for the multi-depot vehicle routing problem. Robotics and Computer- Proceedings of the VI International Logistics and Supply Chain Congress, November
Integrated Manufacturing, 26(6), 564569. 67, Istanbul, Turkey. Available online at: <http://research.sabanciuniv.edu/
Montoya-Torres, J. R. (2014) Designing sustainable supply chains based on the triple 13087/1/SurveyMultiTripVRP.pdf>. Accessed: October 13, 2011.
bottom line approach. Working Paper, Universidad de La Sabana, Chia, Sepehri, M. M., & Kargari, M. (2010). Optimizing the service card in multi-depot,
Colombia. Submitted to Annals of Operations Research. multi-product and multi-level vehicle routing problems with the aim of
Montoya-Torres, J. R., Alfonso-Lizarazo, E, Gutirrez-Franco, E., & Halabi, A. X. minimizing the total distribution costs. International Journal of Industrial
(2009). Using randomization to solve the deterministic vehicle routing problem Engineering and Production Management, 21(1), 4961.
with service time requirements. In M. D. Rossetti, R. R. Hill, B. Johanson, A. Seth, A., Klabjan, D., & Ferreira, P. M. (2013). Analyses of advanced iterated tour
Dunkin, & R. G. Ingalls (Eds.), Proceedings of the 2009 winter simulation partitioning heuristics for generalized vehicle routing problems. Networks,
conference (pp. 29892994). 61(4), 290308.
 J.R. Montoya-Torres et al. / Computers & Industrial Engineering 79 (2015) 115129 129

Sitek, P., Wikarek, J., & Grzybowska, K. (2014). A multi-agent approach to the multi- Villegas, J. G., Prins, C., Prodhon, C., Medaglia, A. L., & Velasco, N. (2010). GRASP/VND
echelon capacitated vehicle routing problem. Highlights of Practical Applications and multi-start evolutionary local search for the single truck and trailer routing
of Heterogeneous Multi-Agent Systems. The PAAMS Collection, Communications in problem with satellite depots. Engineering Applications of Articial Intelligence,
Computer and Information Science, 430, 121132. 23(5), 780794.
Skok, M., Skrlec, D., & Krajcar, S. (2000). The non-xed destination multiple depot Wang, Y. (2013). Research of multi-depot vehicle routing problem by cellular ant
capacitated vehicle routing problem and genetic algorithms. In Proceedings of algorithm. Journal of Computers, 8(7), 17221727.
the 22nd international conference on information technology interfaces (pp. 403 Wang, S., Gao, L., Cui, X., & Chen, X. (2007). Research on Multi-depot Single Vehicle
408). Routing Problem. Control Engineering of China, 2007-06. Online at: <http://
Skok, M., Skrlec, D., & Krajcar, S. (2001). The genetic algorithm of vehicles from en.cnki.com.cn/Article_en/CJFDTOTAL-JZDF200706003.htm>. Accessed 1 May
multiple depots to a number of delivery points. In Proceedings of the 2001 WSES 2012.
international conference on evolutionary computation (EC2001) (pp. 61016108). Wang, Z., Zhang, J., & Wang, X. (2011). A modied variable neighborhood search
Solomon, M. M., & Desrosiers, J. (1988). Time window constrained routing and algorithm for the multi depot vehicle routing problem with time windows.
scheduling problems. Transportation Science, 22(1), 113. Chinese Journal of Management Science, 19(2), 99109.
Sombunthama, P., & Kachitvichyanukulb, V. (2010). Multi-depot vehicle routing Wasner, M., & Zapfel, G. (2004). An integrated multi-depot hub-location vehicle
problem with pickup and delivery requests. AIP Conference Proceedings of the routing model for network planning of parcel service. International Journal of
International MultiConference of Engineers and Computer Scientists, 1285, 7185. Production Economics, 90(3), 403419.
Songyan, C., & Akio, I. (2005). The multi-supplier multi-depot vehicle routing Weise, T., Podlich, A., & Gorldt, C. (2010). Solving real-world vehicle routing
problem. Journal of the Japan Institute of Navigation, 113, 201208. problems with evolutionary algorithms. In R. Chiong, & S. Dhakal (Eds.), Natural
Songyan, C., Akio, I., & Bai, Z. (2005). A SA-based heuristic for the multi-depot intelligence for scheduling, planning and packing problems (pp. 2953).
vehicle routing problem. Journal of Japan Institute of Navigation, 113, 209216. Wilger, T., Maurer, R. (1992) A heuristic-based algorithm for solving multiple supply
Subramanian, A., Uchoa, E., & Ochi, L. S. (2013). A hybrid algorithm for a class of vehicle routing problems (Master of Science thesis). Colorado School of Mines.
vehicle routing problems. Computers & Operations Research, 40(10), 25192531. Wu, T. H., Low, C., & Bai, J. W. (2002). Heuristic solutions to multi-depot location
Surekha, P., & Sumathi, S. (2011). Solution to multi-depot vehicle routing problem routing problems. Computers & Operations Research, 29(10), 13931415.
using genetic algorithms. World Applied Programming, 1(3), 118131. Xu, Y., & Jiang, W. (2014). An improved variable neighborhood search algorithm for
Tan, K. C., Chew, Y. H., & Lee, L. H. (2006). A hybrid multi-objective evolutionary multi-depot heterogeneous vehicle routing problem based on hybrid operators.
algorithm for solving truck and trailer vehicle routing problems. European International Journal of Control and Automation, 7(3), 299316.
Journal of Operational Research, 172(3), 855885. Xu, Y., Wang, L., & Yang, Y. (2012). A new variable neighborhood search algorithm
Tavakkoli-Moghaddam, R., Makui, A., & Mazloomi, Z. (2010). A new integrated for the multi-depot heterogeneous vehicle routing problem with time windows.
mathematical model for a bi-objective multi-depot location-routing problem Electronic Notes in Discrete Mathematics, 39, 289296.
solved by a multi-objective scatter search algorithm. Journal of Manufacturing Yang, Y. (2008). An improved genetic algorithm for multiple-depot and
Systems, 29(34), 111119. heterogeneous-vehicle routing problem. Computer and Modernization, 2008
Thangiah, S. R., & Salhi, S. (2001). Genetic clustering: An adaptive heuristic for the 09. Available online at: <http://en.cnki.com.cn/Article_en/CJFDTOTAL-
multidepot vehicle routing problem. Applied Articial Intelligence, 15(4), JYXH200809002.htm>. Accessed 1 May 2012.
361383. Yang, W. T., & Chu, L. C. (2000). A heuristic algorithm for the multi-depot periodic
Ting, C. J., & Chen, C. H. (2009). Combination of multiple ant colony system and vehicle routing problem. Journal of Information & Optimization Sciences, 22,
simulated annealing for the multidepot vehicle routing problem with time 359367.
windows. Transportation Research Record: Journal of the Transportation Research Yang, Y., Cui, Z., & Cheng, J. (2006). An improved genetic algorithm for multiple-
Board, 2089, 8592. depot vehicle routing problem with time windows. Journal of Soochow
Toth, P., & Vigo, D. (2002). The vehicle routing problem. SIAM monographs on discrete University, 200602. Available at: <http://en.cnki.com.cn/Article_en/CJFDTotal-
mathematics and applications, Vol. 9. Philadelphia: SIAM. SILK200602004.htm>. Accessed 1 May 2012.
Tsirimpas, P., Tatarakis, A., Minis, I., & Kyriakidis, E. G. (2007). Single vehicle routing Yang, H., Zhou, Y., Cui, Z., & He, M. (2011). Vehicle routing problem with multi-
with a predened customer sequence and multiple depot returns. European depot and multi-task. Advances in Information Sciences and Service Sciences, 3(6),
Journal of Operational Research, 187(2), 483495. 320327.
Tzn, D., & Burke, L. (1999). A two-phase tabu search approach to the location Yu, B., Yang, Z. Z., & Xie, J. X. (2011). A parallel improved ant colony optimization for
routing problem. European Journal of Operational Research, 116(1), 8799. multi-depot vehicle routing problem. Journal of the Operational Research Society,
Vahed, A.R., Crainic, T.G., Gendreau, M., & Rei, W. (in press) Fleet-sizing for multi- 62, 183188.
depot and periodic vehicle routing problems using a modular heuristic Ycenur, G. N., & Demirel, N. . (2011a). A new geometric shape-based genetic
algorithm. Computers & Operations Research. doi: http://dx.doi.org/10.1016/ clustering algorithm for the multi-depot vehicle routing problem. Expert
j.cor.2014.07.004. Systems with Applications, 38(9), 1185911865.
Venkatasubbaiah, K., Acharyulu, S. G., Chandra Mouli, K. V. V. (2011) Fuzzy goal Ycenur, G. N., & Demirel, N. . (2011b). A hybrid algorithm with genetic algorithm
programming method for solving multi-objective transportation problems. and ant colony optimization for solving multi-depot vehicle routing problems.
Global Journal of Research in Engineering, 11(3). Available at: <http:// Journal of Engineering and Natural Sciences, 340350.
globaljournals.org/GJRE_Volume11/2-Fuzzy-Goal-Programming-Method-for- Zarandi, M. H. F., Hemmati, A., & Davari, S. (2011). The multi-depot capacitated
Solving.pdf>. Accessed 23 October 2011. location-routing problem with fuzzy travel times. Expert Systems with
Vianna, D. S., Ochi, L. S., & Drummond, L. M. (1999). A parallel hybrid evolutionary Applications, 38(8), 1007510084.
metaheuristic for the period vehicle routing problem. In J. Rolim, F. Mueller, A. Zhang, M., Jiang, D., & Tang, X. (2002). Full load vehicle routing with multiple
Y. Zomaya, F. Ercal, S. Olariu, B. Ravindran, J. Gustafsson, H. Takada, R. Olsson, & depots: new network ow based algorithm. Journal of Systems Engineering,
L. V. Kale et al. (Eds.), Parallel and distributed processing. Lecture notes in computer 17(2), 216220.
science. Vol. 1586 (pp. 183191). Berlin: Springer-Verlag. Zhang, J., Tang, J., & Fung, R. Y. K. (2011). A scatter search for multi-depot vehicle
Vidal, T., Crainic, T. G., Gendreau, M., & Prins, C. (2014). Implicit depot assignments routing problem with weight-related cost. Asia-Pacic Journal of Operational
and rotations in vehicle routing heuristics. European Journal of Operational Research, 28(3), 323348.
Research, 237(1), 1528. Zhen, T., & Zhang, Q. (2009). A combining heuristic algorithm for the multi-depot
Vidal, T., Crainic, T.G., Gendreau, M., Lahrichi, N., & Rei, W. (2010) A hybrid genetic vehicle routing problem with inter-depot routes. In Proceedings of the 2009
algorithm for multi-depot and periodic vehicle routing problems. CIRRELT- International Joint Conference on Articial Intelligence (pp. 436439).
2010-34. Available online at: <https://www.cirrelt.ca/DocumentsTravail/
CIRRELT-2010-34.pdf>. Accessed 23 October 2011.