See discussions, stats, and author profiles for this publication at: [619305]

See discussions, stats, and author profiles for this publication at:
https://www.researchgate.net/publication/306118502
Trends in transportation and logistics
Article

in

European Journal of Operational Research
· August 2016
DOI: 10.1016/j.ejor.2016.08.032
CITATIONS
6
READS
1,134
1 author:
M.Grazia Speranza
Università degli Studi di Brescia
244

PUBLICATIONS

5,340

CITATIONS

SEE PROFILE
All content following this page was uploaded by
M.Grazia Speranza
on 13 October 2017.
The user has requested enhancement of the downloaded file.

T rends in transp ortation and logistics
M. Grazia Sp eranza
Departmen t of Economics and Managemen t
Univ ersit y of Brescia, Italy
Email: grazia.sp [anonimizat]
August 1, 2016
Abstract
Problems in transp ortation and logistics had to b e tac kled long
b efore computers and Op erational Researc h (OR) b ecame a v ailable
to supp ort decision making. After the rst optimization mo dels w ere
dev elop ed, OR has substan tially con tributed in making transp ortation
systems ecien t and companies with complex transp ortation and logis-
tics problems comp etitiv e. Ov er the y ears, tec hnology has ev olv ed and
the same has done OR. In this pap er, the history of problems and OR
con tributions in transp ortation and logistics will b e shortly review ed
together with the ev olution of tec hnology . Then, the future trends in
this area will b e discussed together with p oten tial OR con tributions.
Keyw ords : History , supply c hain managemen t, transp ortation and logistics,
mobilit y , trends, op erational researc h.
1 In tro duction
Problems in transp ortation and logistics had to b e tac kled long b efore com-
puters w ere in v en ted and Op erational Researc h (OR) b ecame a discipline
aimed at dev eloping mo dels and tec hniques to supp ort decision making.
After the rst optimization mo dels w ere dev elop ed, OR has substan tially
con tributed in making transp ortation systems ecien t and companies with
complex transp ortation and logistics problems comp etitiv e. OR, as a sys-
tem scienc e , has captured the complexities of problems and the in teractions
among parts of a system to impro v e the qualit y of decision making. The OR
1

metho ds ha v e b een dep enden t on data a v ailabilit y and in most cases ha v e re-
lied on computers. The a v ailabilit y of more data and of more computational
capacit y ha v e made the OR metho ds more p o w erful.
Ov er the y ears, tec hnology has ev olv ed and the same has done OR. The
In ternet, tec hnological dev elopmen ts, the individual use of information and
comm unication devices, the widespread a v ailabilit y of massiv e amoun ts of
data ha v e created new c hallenges and opp ortunities to transp ortation and
logistic systems, and to researc hers in OR as w ell.
The aim of this pap er is to sho w that, o v er time, the con tributions of
OR to transp ortation and logistics ha v e ev olv ed, follo wing the ev olution of
problems in the area and tec hnology , and that the recen t trends are creating
exciting opp ortunities.
In Section 2 the history of problems and OR con tributions in transp orta-
tion and logistics will b e sk etc hed out, together with the ev olution of tec h-
nology , to outline that problems in the area, tec hnology and OR are strongly
in terconnected. Ma jor general tec hnological phenomena, namely big data
and In ternet of Things, are shortly presen ted in Section 3. Problems in logis-
tics and supply c hain managemen t are mainly related to the transp ortation
of go o ds and are generally treated separately from problems concerning mo-
bilit y , that is the transp ortation of p eople. The trends in logistics and supply
c hain managemen t will b e discussed, together with p oten tial OR con tribu-
tions, in Section 4, whereas the trends in mobilit y problems will b e discussed
in Section 5. Some conclusions will b e nally dra wn in Section 6.
2 The history of transp ortation and logistics
The history of problems in transp ortation and logistics is deep-ro oted. Only
recen tly OR has con tributed to their solution, accompanied b y the birth and
gro wth of computerized systems.
2.1 T ransp ortation and logistics
The history of transp ortation and logistics is as long as the history of mankind,
but has b een mark ed b y recen t milestones. The railroad w as disco v ered at the
b eginning of the nineteen th cen tury , the airplane in 1903. In maritime trans-
p ortation, the in v en tion of the sea con tainer is dated 1956 and has impacted
sea transp ort dramatically . No w ada ys logistics, and the broader concept of
supply c hain managemen t, is mainly in tended as a business function that
has the scop e to mak e go o ds a v ailable where and when needed and in the
needed quan tities. T ransp ortation managemen t can b e seen as part of logis-
2

tics, when referred to the business pro cesses. Ho w ev er, not only go o ds but
also p eople need to b e transp orted. In the old times p eople used to w alk or
tra v el b y horse or carriage or ship. No w ada ys dieren t transp ortation means
are a v ailable with dieren t lev els of cost and comfort.
T raditionally , freigh t transp ortation problems ha v e b een indep enden t of
p eople transp ortation problems. Moreo v er, whereas freigh t transp ortation is
a relev an t problem in the priv ate sector, p eople transp ortation problems ha v e
b een mostly – with the exception of airlines – faced b y the public sector that
is resp onsible for the public transp ortation (mass transit) system and for the
design of the infrastructure for the mo v emen t of priv ate v ehicles. Fleets of
v ehicles, mainly buses and trains, ha v e needed to b e co ordinated in terms of
routes, sc hedules, crew and OR has oered fundamen tal con tributions to the
optimization of these systems.
2.2 Op erational researc h
T ransp ortation and logistics problems ha v e b een studied for a long time b y
researc hers and practitioners in op erational researc h. In fact, the rst con-
tribution dates bac k to 1930 and is due to T olsto, as rep orted b y Sc hrijv er
[38]. It is a mo del for the solution of a practical problem related to the
transp ortation of salt, cemen t, and other cargo b et w een sources and destina-
tions along the railw a y net w ork of the So viet Union. The author studied the
transp ortation problem and describ ed v arious solution approac hes and the
no w w ell-kno wn idea that an optimal solution do es not ha v e an y negativ e-
cost cycle in its residual graph. As men tioned in [38], T olsto migh t ha v e
b een the rst to observ e that the cycle condition is necessary for optimalit y .
A large – for that time – size instance with 10 origins and 68 destinations
w as solv ed taking adv an tage of the structure of the railw a y net w ork. Later,
the transp ortation problem w as form ulated b y Hitc hco c k in 1941 [23], and a
cycle criterion for optimalit y w as considered b y Kan toro vic h in 1942 [26].
The history of the con tributions of OR to transp ortation and logistics
follo w ed the ev olution of the problems in the area and at the same time
the dev elopmen ts in the information and comm unication tec hnologies (ICT).
The ma jor phases of this history can b e sk etc hed as:
1960s and 70s: T r ansp ortation scienc e emerged. T r ansp ortation mean t traf-
c and public transp ortation, whereas lo gistics w as a y oung eld that
referred to ph ysical distribution and in v en tory managemen t. In the
same p erio d, dieren t programming languages w ere dev elop ed. The
rst F OR TRAN compiler w as deliv ered, in fact, in 1957. In the six-
ties 40 F OR TRAN compilers b ecame a v ailable. F OR TRAN w as de-
3

v elop ed for scien tic and engineering applications and dominated this
area of programming for o v er half a cen tury . Besides F OR TRAN, other
languages w ere dev elop ed: in 1968 Logo, in 1970 P ascal, in 1972 C,
Smalltalk, Prolog, in 1978 SQL;
1980s: It is the p erio d of the study of truc king (common carriers and pri-
v ate eets). In this decade rail and sea transp ortation emerged. Air
transp ortation also emerged as a distinct researc h area. During the
early 1980s, home computers w ere dev elop ed for household use, with
soft w are for p ersonal pro ductivit y , programming and games;
1990s: T r ansp ortation included passenger and freigh t transp ortation. L o-
gistics dev elop ed with a fo cus on op erations and shipp ers in to supply
chain management . T r ansp ortation and lo gistics emerged to co v er a
broader v ariet y of problems. Since the mid-1990s, the In ternet has had
a rev olutionary impact on culture and commerce, including the rise of
near-instan t comm unication b y electronic mail, instan t messaging, and
the W orld Wide W eb with its discussion forums, blogs, so cial net w ork-
ing, and online shopping sites;
2000/2010: T r ansp ortation and lo gistics co v ers a con tin uously gro wing n um-
b er of applications. The traditional barrier b et w een freigh t and passen-
ger transp ortation seems to b ecome thinner and thinner. The mobile
apps (that stands for applications) are made a v ailable through digi-
tal distribution platform to mobile devices suc h as smartphones and
tablets.
The researc h in transp ortation and logistics has not only pro duced an
adv ance of kno wledge, with academic results measurable in pap ers published
and conferences organized. Ha ving alw a ys b een driv en b y real problems, the
researc h has pro duced mo dels and algorithms that ha v e b een em b edded in
soft w are pac k ages, used b y companies in the priv ate and public sectors. T w o
surv eys published in OR/MS T o da y , one on soft w are for supply c hain man-
agemen t [4], and one sp ecically for v ehicle routing problems [22], witness
this impact. The t w o surv eys are accompanied b y a summary of the prob-
lems addressed in the soft w are and the corresp onding OR to ols, from facilit y
lo cation to w arehouse managemen t, from lot sizing to pro duction sc hedul-
ing, from supply c hain net w ork design to in v en tory managemen t, from eet
managemen t to v ehicle routing (see also the recen t b o oks [41] and [15]).
4

3 Big data and In ternet of Things
The most recen t tec hnological adv ances are related to the explosion of digital
data, the so called big data , and to the expansion of the concept of In ternet
to the so called In ternet of Things (IoT), also called the In ternet of ob jects.
The n um b er of searc hes in Go ogle for `big data' has exceeded in 2013
the n um b er of searc hes for `supply c hain managemen t' (see [42]). This do es
not imply that data is more imp ortan t than supply c hain managemen t but
certainly is a signal of the gro wing p erception that the a v ailabilit y of massiv e
quan tities of data is relev an t to businesses and to services, to the priv ate and
to the public sectors, to companies and to institutions.
What mak es big data dieren t from traditional data? In [31] three main
dierences are iden tied: v olume, v elo cit y and v ariet y . `More data cross the
in ternet ev ery second than w ere stored in the en tire in ternet just 20 y ears
ago' [31]. The sp eed of data creation is for man y applications extremely im-
p ortan t, p ossibly more imp ortan t than the v olume of data. Big data tak es a
v ariet y of forms, from messages to images, global p ositioning system (GPS)
signals from cell phones, readings from sensors. The so cial net w orks, smart-
phones and mobile devices are sources of big data and pro vide enormous
amoun ts of data related to p eople, activities, lo cations. Smartphones and
mobile devices ha v e b ecome so ubiquitous that it is easy to forget that they
did not exist less than ten y ears ago. Because of big data, managers and
decision mak ers can kno w more and transform that kno wledge in to impro v ed
decision making and p erformance.
Data-driv en decisions are b etter decisions. This is an ob vious concept
to op erational researc hers. Ho w ev er, the big data is creating a new broad
range of opp ortunities for op erational researc hers, in particular in the eld
of transp ortation and logistics. In fact, while big data is an extremely p op-
ular expression and data scien tists are requested b y an increasing n um b er
of companies (see [17]), acquiring big data is far from b eing sucien t. Big
data is the input to adv anced quan titativ e to ols that can lead companies and
institutions to b etter decisions (see [11]).
The big data phenomenon is related to, and also partially caused b y , the
IoT (see, for example, [9] and [21]) whic h is the net w ork of ph ysical ob jects
or `things' em b edded with electronics, soft w are, sensors, and net w ork con-
nectivit y , whic h enables these ob jects to collect and exc hange data. The
IoT allo ws ob jects to b e sensed and con trolled remotely across an existing
net w ork infrastructure, creating opp ortunities for more direct in tegration b e-
t w een the ph ysical w orld and computer-based systems. The IoT encompasses
smart grids, smart homes, in telligen t transp ortation and smart cities. In the
IoT eac h thing is iden tied b y its em b edded tec hnology and is able to in ter-
5

op erate within the existing In ternet infrastructure. The IoT will increase the
ubiquit y of the In ternet and lead to a highly distributed net w ork of devices
comm unicating with h uman b eings as w ell as with other devices. The IoT
is exp ected to generate in the future larger amoun ts of data than a v ailable
no w ada ys.
Big data and IoT are op ening tremendous opp ortunities for a large n um-
b er of no v el applications and researc h pro jects.
4 T rends in supply c hain managemen t
The aim of this section is to outline new trends and researc h opp ortunities
in supply c hain managemen t. The en vironmen t for supply c hain activities
is c hanging and exp erts predict that man y c hanges will o ccur in the near
future. In [40] ten trends w ere iden tied as inuen tial in the near future
of the supply c hain managemen t industry . W e shortly summarize here the
trends that are most relev an t to OR:
 Systemic fo cus: Optimization of the en tire supply c hain net w ork, cus-
tomer v alue co-creation;
 Information syn thesis: Information is holistically shared, join t in ter-
pretation to impro v e p erformance;
 Collab orativ e relationships: Join t accoun tabilit y and rew ards, total
system v alue creation;
 Demand shaping: Proactiv ely inuencing demand, total system v alue
creation;
 T ransformational agilit y: Constan tly c hanging conditions;
 Flexible net w ork in tegration: Dynamic selection of partners upstream
and do wnstream;
 Global optimization.
Three ma jor directions for researc h can b e deriv ed from these trends: a
systemic, a collab orativ e and a dynamic direction. In the follo wing w e discuss
the researc h opp ortunities asso ciated with eac h of these.
6

4.1 Systemic direction
Op erational researc h has con tributed to decision making in sev eral areas
of supply c hain managemen t. The systemic direction suggests that b etter
solutions to problems can b e iden tied when broader parts of the supply
c hain are join tly mo deled and optimized. In fact, sev eral researc h eorts
ha v e already b een made in recen t y ears in this direction.
As an example, in the area of v ehicle routing, sev eral pap ers ha v e studied
more global problems with resp ect to the classical routing problems aimed
at nding the routes of v ehicles only , giv en lo cations, demands of customers,
time windo ws. In tegrated v ehicle routing problems is the expression increas-
ingly used to denote the class of problems where the routing decisions are
tac kled together with other decisions (as outlined b y the sp ecial issue edited
b y Bekta³ et al. [12]). Lo cation-routing problems join tly optimize lo cation
and routing. In v en tory-routing problems com bine routing and in v en tory
managemen t. Pro duction-routing problems in tegrate pro duction, routing,
and usually also in v en tory decisions. Multi-ec helon routing problems opti-
mize the routes of v ehicles in distribution systems comprising t w o or more
ec helons. Routing problems with loading constrain ts sim ultaneously optimize
the routing of v ehicles and the loading of go o ds on them.
In tegrated v ehicle routing problems com bine problems that are usually
NP- hard b y themselv es (routing problems are among the hardest com binato-
rial problems). Ho w ev er, solving the problems indep enden tly , ev en b y means
of exact metho ds, leads to a sub-optimal solution for the in tegrated problem.
One of the rst pap ers that sho w ed the b enets of in tegrated decisions is
due to Chandra and Fisher [14]. More recen tly , Arc hetti and Sp eranza [6]
compared the heuristic solution of an in v en tory-routing problem with the so-
lution obtained b y sequen tially and optimally solving the in v en tory manage-
men t and the routing problems. The sequen tial solution mo dels a traditional
managemen t st yle of a supply c hain where customers con trol their optimal
in v en tory managemen t p olicy and decide order times and quan tities. Only
afterw ards, the supplier organizes an optimal distribution that, ho w ev er, has
to tak e customer times and quan tities as constrain ts. The in v en tory-routing
problem mo dels instead a more recen t in tegrated managemen t p olicy , called
V endor Managed In v en tory (VMI), where the supplier is resp onsible for the
distribution as w ell as the in v en tory at its customers (see, for example, the
review of VMI b y Marquès et al. [28]). The researc h direction to w ards more
in tegrated problems is consisten t with the trends sk etc hed ab o v e. More in-
tegrated optimization problems mo del more in tegrated managemen t st yles
of supply c hains, con tribute to exploit the adv an tages of the in tegration and
can quan tify the b enets. In [6] the results of computational tests sho w that
7

solutions of the in v en tory-routing problem allo w a v erage sa vings of 10%, with
a v erage sa vings on in v en tory and transp ortation costs of 15% and 9%, resp ec-
tiv ely . Th us, if a heuristic is used for the solution of the in tegrated problem
that generates an a v erage error of less than 10%, the in tegration oers b ene-
ts with resp ect to the sequen tial solution, ev en if optimal, of the individual
problems.
4.2 Collab orativ e direction
The trends include collab orativ e relationships. Collab oration in supply c hain
managemen t has b een widely discussed (see, for example, [10], [24] and [19])
and man y strategies ha v e b een suggested, including, among the most p opu-
lar ones, VMI and Collab orativ e Planning, F orecasting and Replenishmen t
(CPFR) initiativ es. Collab oration can b e seen as a to ol that enables in tegra-
tion and global optimization of a supply c hain. It is b ey ond the scop e of this
pap er to explore the complexities of the implemen tation of collab oration ini-
tiativ es and discuss when and wh y collab oration can b e eectiv e in practice.
The goal here is to start from the observ ation that collab oration is a trend
in supply c hain managemen t, enabled b y the tec hnology and stim ulated b y
increased comp etition and exp ected b enets, and to argue that new opti-
mization problems arise when decision making tak es place in a collab orativ e
en vironmen t. Collab oration initiativ es ma y fail for sev eral reasons and the
lac k of exploitation of the p oten tial b enets is one of those reasons. Here is
where OR can con tribute.
The c hallenges of in tegrating in ternal and external op erations are kno wn
(see, for example, [24]). In this section for collab oration w e in tend external
collab oration, that is collab oration with companies that are external to the
supply c hain. Collab oration inside the supply c hain can b e seen as falling in
the systemic direction.
P artners of a collab oration initiativ e decide to w ork together b ecause
they exp ect to impro v e the p erformance of their o wn business through col-
lab oration. Whereas collab oration will c hange their b eha viour and imply
in teractions among partners in a join t eort to w ards in tegration, eac h part-
ner will b e fo cused on its o wn business more than on a global p erformance.
Th us, in tegration m ust b e mediated with individual in terests to mak e the
collab oration initiativ e successful. This essen tial concept in collab oration
ma y mak e mo dels for decision supp ort in collab oration initiativ es dieren t
from mo dels for global optimization.
As an example, let us consider collab oration among carriers. Statistics
sho w (see [1]) that appro ximately 90% of freigh t tra v els on road and that in
all Europ ean coun tries the p ercen tage of empt y truc ks tra v eling on road and
8

con tributing to trac, p ollution, acciden ts is b et w een 15% and 30%. The
a v erage load of a truc k is m uc h lo w er than its capacit y and particularly lo w in
cit y distribution. The n um b er of truc ks on road is m uc h higher than it should
b e. Sev eral causes of these negativ e statistics can b e iden tied. Among
these, w e certainly ha v e the size of carriers, the disp ersion of customers, the
short lead time b et w een order and deliv ery times, caused in particular b y
the increasing v olumes generated b y e-commerce activities. Collab oration
among carriers ma y impro v e the statistics and generate economic b enets
for the carriers in v olv ed as w ell as so cial and en vironmen tal b enets.
T o illustrate the concept, let us consider the problem studied b y F ernán-
dez et al. [20], where a collab oration sc heme is adopted b y a group of carriers.
Eac h carrier ma y decide to serv e a subset of its customers and to share the
other customers with other carriers. If a customer is shared it ma y b e serv ed
b y an y of the carriers of the group. A carrier will c ho ose to serv e customers
with high v olumes or con v enien tly lo cated or v aluable for an y other reason.
A customer with lo w demand, and far from the dep ot and from other cus-
tomers will b e lik ely shared. The prot coming from a shared customer will
b e partly collected b y the carrier `o wning' the customer and partly b y the
carrier actually serving the customer. The rev en ue sharing agreemen t is part
of the collab oration sc heme. In the solution where the sum of the costs of
all carriers is minimized, eac h carrier will serv e its customers, some of the
customers shared b y other carriers, and p ossibly some of its o wn shared cus-
tomers that ma y b ecome con v enien t when com bined with customers shared
b y others. This solution will generate sa vings with resp ect to the total cost of
the solution where carriers do not collab orate. Ho w ev er, suc h solution ma y b e
suc h that the prot of a carrier is lo w er than its prot without collab oration.
Suc h a situation is lik ely to b e unacceptable to a carrier, esp ecially if not
exp erienced only o ccasionally , and ma y mak e the collab oration fail. In [20]
a mo del is, th us, suggested where the prot of eac h carrier is constrained to
b e not lo w er than the prot that w ould b e gained without collab oration. In
this w a y , the collab orativ e solution will b e b enecial to eac h individual car-
rier. Computational results sho w that the total prot increase in a solution
where carriers collab orate, with the individual prot guaran tee, is on a v er-
age 7% higher than the total prot without collab oration. The results also
sho w that the prot increase strongly dep ends on the lo cation and demand
of customers, ranging from small p ositiv e v alues to up to 85%.
4.3 Dynamic direction
The transformational agilit y and the constan tly c hanging conditions listed as
trends are caused b y the con tin uously c hanging o w of data ab out customers,
9

purc hases, deliv eries, lo cations, in v en tories. This in turn mak es the problems
in trinsically m uc h more dynamic than they used to b e. Systems should
b ecome more reactiv e to c hanges and pro vide more eectiv e resp onses to
customers whose demand is b ecoming more and more v ariable o v er time, due
to the increasing v olumes of e-commerce. This latter trend mak es the demand
also dicult to predict. As planning activities based on forecasting will
remain essen tial in supply c hain managemen t, esp ecially in the upp er parts
of the supply c hains (see [39]), mo dels should also capture all the p ossible,
uncertain, information a v ailable on future outcomes.
Most classical optimization mo dels assume that all relev an t information
is a v ailable at the momen t a mo del is built, that the mo del is then run and
the solution obtained en tirely implemen ted. This w as a realistic assumption
in a w orld where it w as v ery time consuming and costly to collect data and
where the data w ere up dated rarely . This assumption is b ecoming less and
less acceptable, b ecause solutions need to b e revised shortly , b efore b eing
completely implemen ted. Although dynamic problems in transp ortation ha v e
b een discussed for a long time (for example, b y Psaraftis [35]), researc h on
dynamic and sto c hastic v ehicle routing problems receiv ed increasing in terest
only in the last decade (see the recen t surv ey b y Ritzinger et al. [36]). In
fact, in [36] the imp ortance of appropriately mo deling dynamic ev en ts and
sim ultaneously incorp orating information ab out the uncertain t y of future
ev en ts is outlined.
Sev eral researc h opp ortunities related to the dynamic direction in supply
c hain managemen t are discussed in [42]. Moreo v er, the data is con tin uously
c hanging and long computational times b ecome less and less acceptable. Ho w
m uc h computational eort is w orth in v esting in a mo del whose solution will
only b e partially implemen ted? When a c hange in the data should imply a
rerun of a mo del? Or what c hanges in the data mak e the rerunning of a mo del
b enecial? The need to run a mo del frequen tly b ecause of the con tin uous
up date of data creates other relev an t researc h issues. Can w e tak e adv an tage
of the w ork done b y an algorithm for the solution of an optimization problem
to sp eed up the solution of the next problem, where some data, but not all,
ha v e c hanged?
5 T rends in transp ortation
Although priv ate cars remain the dominan t transp ortation mo de for the large
ma jorit y of p eople, the set of mobilit y options is gro wing. Startups in this
sector establish themselv es within a short time. Ub er, Grabtaxi, BlaBlaCar,
Zip car are only some of the names corresp onding to companies that oer an
10

alternativ e transp ortation mo de to p eople, some for short, others for long
distances. Y oung p eople tend to use these new options and to dela y the
purc hase of a car and the acquisition of a driving licence.
In [34] six ma jor trends in p eople transp ortation are presen ted that will
c hange the w a y w e mo v e:
 Autonomous v ehicles: Hands-free and feet-free driving is realit y , fully
autonomous v ehicles will b ecome realit y shortly;
 Electric v ehicles: Mainly transit buses and short-range v ehicles are
electric at presen t, electric v ehicles are b ecoming more economical and
can tra v el longer without b eing c harged;
 Connected v ehicles: T rac data are b ecoming a v ailable on v ehicles,
v ehicles are equipp ed with In ternet connectivit y;
 Collab orativ e consumption: On-demand mobilit y options are gro wing,
collab orativ e options enable mobilit y without mostly un used individual
cars;
 Ecien t m ulti-mo dal net w orks: Cro wdsourcing transit data will adapt
sc hedules to tra v ellers needs, m ultiple trip options will b e oered to
tra v ellers;
 New materials: Ligh ter v ehicles will b e designed, also to increase the
distance tra v eled b y electric v ehicles.
It ma y not tak e long to see a eet of autonomous, shared v ehicles, con-
nected to the road infrastructure, to the In ternet, and to a broader net w ork
of public transit options. In the rest of this section some researc h directions
will b e discussed.
5.1 Electric v ehicles
Hybrid electric v ehicles, battery electric v ehicles, plug-in h ybrid electric v ehi-
cles from b eing exotic w ords ha v e b ecome part of the options of an y p oten tial
car buy er no w ada ys. W e will refer to all these classes of v ehicles as electric
v ehicles. The duration of batteries has increased and th us the autonom y
of v ehicles. Charging stations, though still rare, are increasing in n um b er.
Costs are still high but are exp ected to decrease. Incen tiv es to the use of
electric v ehicles come from p olitical institutions that see the p ositiv e impact
coming from the c hange to the en vironmen t. While the global impact of a
11

massiv e substitution of traditional with electric v ehicles remains to b e as-
sessed, esp ecially in terms of pro duction of electricit y , the trend to w ards the
use of electric v ehicles seems to b e irrev ersible. This will certainly lead to
en vironmen tal b enets, esp ecially in densely p opulated areas.
Sev eral pap ers ha v e already app eared that tac kle optimization problems
sp ecically arising for electric v ehicles (see, for example, [2], [37], [33], [43]).
Dep ending on the tec hnological c haracteristics of an electric v ehicle, dier-
en t problems b ecome relev an t. Suc h problems ma y concern the lo cation of
c harging stations, the routing of v ehicles constrained b y the limited auton-
om y and the rarit y of the c harging stations, the reserv ation of a battery , the
eet managemen t.
5.2 Reduction of tra v eling v ehicles and parking space
T rac congestion is a dramatic problem in ev ery coun try . P eople ha v e to
queue in their cars daily to reac h their w orking place, to tak e c hildren to
sc ho ol, to p erform an y regular activit y . Green areas are transformed in park-
ing spaces. Queueing is not an exceptional ev en t but rather, esp ecially in
urban areas, a regular ev en t that causes dela ys and stress. Dela ys in turn
ha v e h uge economic and so cial consequences. The substitution of traditional
v ehicles with electric v ehicles will not reduce the n um b er of tra v eling v ehicles,
the need of parking space or the congestion problems.
The n um b er of tra v eling v ehicles can b e reduced only b y reducing the
n um b er of p eople in need of tra v el and/or b y increasing the n um b er of p eople
transp orted in the same v ehicle. While w e can hardly con tribute to the former
option, our con tribution ma y b e relev an t in supp orting the latter.
One of the main reasons that leads p eople to use their o wn priv ate v ehi-
cles is the lac k of exibilit y of mass transit systems. Suc h mobilit y systems
t ypically w ork on xed itineraries and xed sc hedules. In most cases the fre-
quency is to o lo w and the tra v el time is to o high. Suc h c haracteristics mak e
these systems inappropriate for a transp ortation demand that is extremely
disp ersed in space and time, and requests quic k resp onse and short tra v eling
time.
Demand Resp onsiv e T ransit (DR T) systems (also called dial-a-ride sys-
tems) are exible services that pro vide `do or-to-do or' transp ortation. DR T
systems are no w ada ys mainly implemen ted as services for small groups of
p eople but ha v e attracted a lot of researc h (see [16] for a surv ey and, as
examples of more recen t con tributions, [30] and [27]).
Martínez et al [29] ha v e suggested a classication of DR T systems:
 with xed itineraries and stops, with pre-b o oking;
12

 with xed itineraries and stops with p ossible detours;
 with unsp ecied itineraries and predened stops;
 with unsp ecied itineraries and unsp ecied stops.
The last t yp e of service, whic h is the most exible one, can b e considered
as the closest to the concept of shared taxis.
DR T systems are attracting more and more in terest and DR T service
pro viders b ecome in terested in impro ving the eciency of their op erations.
An implemen tation of a DR T system in Maryland is presen ted b y Mark o vi¢
et al. [27] and the b enets, with 450 trip requests daily , of a computerized
routing and sc heduling system are estimated with ann ual sa vings of $0.82
million, or ab out 18% of the total ann ual exp ense, with resp ect to man ual
op erations.
In [7] a sim ulation study is p erformed where a con v en tional mass transit
system, sa y a system of buses, is oered together with an on-demand service
without xed itineraries and sc hedules that allo ws users to comm unicate the
desired departure time, origin and destination of the trip. The on-demand
service is pro vided through minibuses. A minibus, if acceptable to the user
in terms of arriv al time to destination, will pro vide the service pic king up
the user at the origin of the trip and deliv ering him/her to the destination.
In case neither the con v en tional bus nor the on-demand minibus pro vide an
acceptable service to the user, he/she will use a priv ate car. The analysis
p erformed suggests that the on-demand service w ould dominate the con v en-
tional buses, in terms of n um b er of trips attracted, tra v el time and cost,
that it w ould attract most of the p eople using at presen t a priv ate car, and
it w ould b e more en vironmen tally friendly , as it w ould reduce trac and
congestion.
More recen tly , DR T systems ha v e b een referred to as Flexible T rans-
p ortation Services (FTS) (see [32]) when used as feeder systems for more
traditional public transp ortation services suc h as buses or trains. A recen t
pap er [8] in tro duced a Flexible Mobilit y On Demand (FMOD) system that
oers dieren t services, taxis, shared taxis and minibuses, where the minibus
service w orks as a regular bus service with xed sc hedules.
Man y researc h issues arise in DR T systems, from eet managemen t to dy-
namic routing of v ehicles. Also, when used as feeder systems, sync hronization
problems should b e addressed.
The so called dynamic ride-share systems share with the DR T systems
the goal of increasing the n um b er of p eople sharing the same v ehicles. Suc h
systems aim to bring together tra v elers with similar itineraries and time
sc hedules on short-notice. Eectiv e and ecien t optimization metho ds that
13

matc h driv ers and riders in real-time are necessary for a successful implemen-
tation of suc h systems (see [3] for a review of dynamic ride-sharing systems).
Whereas DR T systems aim at reducing the n um b er of tra v eling v ehicles,
the need of parking space can also b e reduced through car sharing systems.
In a car sharing service a car is pre-b o ok ed, used and returned to a parking
station. One-w a y , with resp ect to t w o-w a y , systems pro vide more exibilit y
to users since cars can b e dropp ed-o at an y station. As recen t pap ers on
optimization mo dels for car sharing problems w e refer to [13] and [18]. Re-
searc h opp ortunities include the lo cation of stations and cars, car relo cation
problems, co ordination of reserv ations.
5.3 Reduction of congestion
W e tend to think that congestion is uniquely determined b y the n um b er of
v ehicles on roads. In fact, this is only partially true, b ecause congestion is
determined also b y the paths follo w ed b y tra v eling v ehicles and b y the time at
whic h v ehicles tra v el. Congestion happ ens when man y v ehicles tra v el along
the same road at the same time. With the tec hnology a v ailable no w ada ys it
b ecomes p ossible, for a giv en n um b er of tra v eling v ehicles with giv en origins
and destinations, to co ordinate tra v eling paths and times.
The most common in-v ehicle device aimed at supp orting driv ers in path
selection is based on a digitalized road net w ork map and a GPS aerial. Giv en
a destination, the na vigation system pro vides an optimal route usually in
terms of distance or tra v el time. Recen tly , na vigation systems p ossess some
real-time trac data and ma y reroute the driv ers to non-congested paths.
Ho w ev er, these systems do not consider the systemic impact of the pro vided
directions. The na vigation devices oer driv ers with close origins and des-
tinations the same information and, as a consequence, route guidance ma y
simply shifts congestion to other roads. The p oten tial for co ordination and
congestion reduction is enormous.
Optimization mo dels ha v e b een presen ted in [25] and [5] with the goal of
nding a system-optimal trac distribution that ensures fairness, that is do es
not increase the shortest paths of driv ers b y more than a giv en p ercen tage.
In b oth pap ers a road net w ork is giv en together with an origin-destination
(OD) matrix that, for eac h OD pair, giv es the o w, that is the n um b er of
cars tra v eling from that origin to that destination in a giv en time p erio d. In
[25] the arc tra v el time is mo deled as a function of the n um b er of v ehicles on
that arc. The resulting mo del is non-linear and a column generation solution
metho d is prop osed. In [5] the arc tra v el time is giv en and the resulting
mo dels are linear. The prop osed approac h assigns paths to driv ers with
the ob jectiv e of minimizing congestion while not increasing their tra v elled
14

distance b y more than the giv en acceptable p ercen tage.
Researc h directions include the dynamic generation of acceptable paths,
time-dep enden t optimization mo dels (b ecause the OD matrix c hanges o v er
time), ev aluation of driv ers b eha viour, impact of incen tiv es to driv ers to
follo w the directions of the system. Previously men tioned tec hnological ad-
v ances, in particular the autonomous or driv erless v ehicles, ma y c hange the
situation and w ould ha v e sev eral b enets in terms of congestion. Driv ers will
get used to trusting their v ehicle and will more lik ely accept a route that is
longer than the shortest one.
6 Conclusions
Recen t tec hnological and automotiv e adv ances are rapidly c hanging the w a y
supply c hains are managed and go o ds and p eople are transp orted. Economic
pressure pushes companies to b ecome more ecien t and eectiv e b y also tak-
ing adv an tage of the tec hnological adv ances. A t the same time, institutions
are driv en b y the sustainabilit y goal, in tended as the abilit y to meet the
needs of the presen t without compromising the abilit y of future generations
to meet their needs. It is exp ected that the h uge economic impact of logistic
costs on companies and of transp ortation on the en vironmen t, together with
new arising business opp ortunities will rapidly c hange transp ortation and
logistics.
Op erational researc h has giv en fundamen tal con tributions to supply c hain
managemen t and transp ortation problems and more essen tial con tributions
are exp ected in resp onse to the new researc h c hallenges. This pap er has
summarized some of the ma jor trends but sev eral others are b ehind the
corner. Opp ortunities for the consolidation of go o ds and p eople on the same
v ehicle are already arising. F or example, customers ma y deliv er go o ds of
other customers, for a small economic incen tiv e. Rev en ue managemen t is
another researc h area that w as only marginally men tioned in this pap er.
Op erational researc h is more vital than ev er and can add relev an t v alue to
new a v ailable tec hnology .
A c kno wledgemen ts
The con ten t of this pap er is inspired b y the ideas presen ted during the plenary
lecture I ga v e at the EUR O conference of Glasgo w in 2015. I am grateful to
ha v e b een giv en the opp ortunit y to reect on the trends in transp ortation
and logistics and to the man y colleagues with whom I w ork and in teract for
15

ha ving con tributed to dev eloping these concepts and ideas.
I also wish to ac kno wledge the suggestions of t w o anon ymous review ers
whic h help ed me impro v e a previous v ersion of this pap er.
References
[1] Il piano nazionale della logistica 2011/2020. Minister o del le Infr astrut-
tur e e dei T r asp orti, Consulta Gener ale p er l'A utostr asp orto e la L o gis-
tic a , Decem b er 2010.
[2] J.D. A dler and P .B. Mirc handani. Online routing and battery reser-
v ations for electric v ehicles with sw appable batteries. T r ansp ortation
R ese ar ch Part B: Metho dolo gic al , 70:285302, 2014.
[3] N. Agatz, A. Erera, M. Sa v elsb ergh, and X. W ang. Optimization for
dynamic ride-sharing: A review. Eur op e an Journal of Op er ational R e-
se ar ch , 223:295303, 2012.
[4] Y. Akso y and A. Derb ez. Soft w are surv ey: Supply c hain managemen t.
OR/MS T o day , June 2003.
[5] E. Angelelli, I. Arsik, V. Morandi, M. Sa v elsb ergh, and M.G. Sp eranza.
Proactiv e route guidance to a v oid congestion. submitte d , 2015.
[6] C. Arc hetti and M.G. Sp eranza. The in v en tory routing problem: the
v alue of in tegration. International T r ansactions in Op er ational R ese ar ch ,
23:393407, 2016.
[7] C. Arc hetti, M.G. Sp eranza, and D. W eyland. On-demand public trans-
p ortation. submitte d , 2015.
[8] B. A taso y , T. Ik eda, X. Song, and M.E. Ben-Akiv a. The concept and
impact analysis of a exible mobilit y on demand system. T r ansp ortation
R ese ar ch Part C: Emer ging T e chnolo gies , 56:373392, 2015.
[9] L. A tzori, A. Iera, and G. Morabito. The in ternet of things: A surv ey .
Computer Networks , 54:27872805, 2010.
[10] M. Barratt. Understanding the meaning of collab oration in the supply
c hain. Supply Chain Management: A n International Journal , 9:3042,
2004.
[11] D. Barton and D. Court. Making adv anced analytics w ork for y ou.
Harvar d Business R eview , 90:7983, 2012.
16

[12] T. Bekta³, G. Lap orte, and D. Vigo. In tegrated v ehicle routing problems.
Computers & Op er ations R ese ar ch , 55:126, 2015.
[13] B. Bo y ac, K.G. Zografos, and N. Geroliminis. An optimization frame-
w ork for the dev elopmen t of ecien t one-w a y car-sharing systems. Eu-
r op e an Journal of Op er ational R ese ar ch , 240:718733, 2015.
[14] P . Chandra and M.L. Fisher. Co ordination of pro duction and distribu-
tion planning. Eur op e an Journal of Op er ational R ese ar ch , 72:503517,
1994.
[15] Á. Corb erán and G. Lap orte. A r c r outing: pr oblems, metho ds, and
applic ations , v olume 20. SIAM, 2015.
[16] J.-F. Cordeau and G. Lap orte. The dial-a-ride problem: mo dels and
algorithms. A nnals of Op er ations R ese ar ch , 153:2946, 2007.
[17] T.H. Da v enp ort and D.J. P atil. Data scien tist: The sexiest job of the
21st cen tury . Harvar d Business R eview , 90:7076, 2012.
[18] G.H. de Almeida Correia and A.P . An tunes. Optimization approac h to
dep ot lo cation and trip selection in one-w a y carsharing systems. T r ans-
p ortation R ese ar ch Part E: L o gistics and T r ansp ortation R eview , 48:233
247, 2012.
[19] S.E. F a w cett, A.M. F a w cett, B.J. W atson, and G.M. Magnan. P eeking
inside the blac k b o x: to w ard an understanding of supply c hain collab o-
ration dynamics. Journal of Supply Chain Management , 48:4472, 2012.
[20] E. F ernández, D. F on tana, and M.G. Sp eranza. On the collab oration
uncapacitated arc routing problem. Computers & Op er ations R ese ar ch ,
67:120131, 2016.
[21] J. Gubbi, R. Buyy a, S. Marusic, and M. P alanisw ami. In ternet of things
(iot): A vision, arc hitectural elemen ts, and future directions. F utur e
Gener ation Computer Systems , 29:16451660, 2013.
[22] R. Hall and J. P art yk a. V ehicle routing soft w are surv ey: Higher exp ec-
tations driv e transformation. OR/MS T o day , F ebruary 2016.
[23] F.L. Hitc hco c k. The distribution of a pro duct from sev eral sources to
n umerous lo calities. Journal of Mathematic al Physics , 20:224230, 1941.
17

[24] M. Holw eg, S. Disney , J. Holmström, and J. Småros. Supply c hain col-
lab oration:: Making sense of the strategy con tin uum. Eur op e an Man-
agement Journal , 23:170181, 2005.
[25] O. Jahn, R. H. Möhring, A. S. Sc h ulz, and N. Stier-Moses. System-
optimal routing of trac o ws with user constrain ts in net w orks with
congestion. Op er ations R ese ar ch , 53:600616, 2005.
[26] L.V. Kan toro vic h. On the translo cation of masses. In Doklady A kademii
Nauk SSSR , v olume 37, pages 199201, 1942.
[27] N. Mark o vi¢, R. Nair, P . Sc honfeld, E. Miller-Ho oks, and M. Mohebbi.
Optimizing dial-a-ride services in maryland: Benets of computerized
routing and sc heduling. T r ansp ortation R ese ar ch Part C: Emer ging
T e chnolo gies , 55:156165, 2015.
[28] G. Marquès, C. Thierry , J. Lamothe, and D. Gourc. A review of v endor
managed in v en tory (VMI): F rom concept to pro cesses. Pr o duction Plan-
ning and Contr ol: The Management of Op er ations , 21:547561, 2010.
[29] L.M. Martínez, J.M. Viegas, and T. Eiró. F orm ulating a new express
minibus service design problem as a clustering problem. T r ansp ortation
Scienc e , 49:8598, 2014.
[30] R. Masson, F. Leh uédé, and O. Péton. The dial-a-ride problem with
transfers. Computers & Op er ations R ese ar ch , 41:1223, 2014.
[31] A. McAfee and E. Brynjolfsson. Big data: The managemen t rev olution.
Harvar d Business R eview , 90:6068, 2012.
[32] C. Mulley and J.D. Nelson. Flexible transp ort services: A new mar-
k et opp ortunit y for public transp ort. R ese ar ch in T r ansp ortation Ec o-
nomics , 25:3945, 2009.
[33] S. P elletier, O. Jabali, and G. Lap orte. Go o ds distribution with electric
v ehicles: Review and researc h p ersp ectiv es. T ec hnical rep ort, T ec hnical
Rep ort CIRREL T-2014-44, CIRREL T, Mon tréal, Canada, 2014.
[34] B. P orter, M. Linse, and Z. Barasz. Six transp ortation trends that will
c hange ho w w e mo v e. F orb es , Jan uary 2015.
[35] H.N. Psaraftis. Dynamic v ehicle routing: Status and prosp ects. A nnals
of Op er ations R ese ar ch , 61:143164, 1995.
18

[36] U. Ritzinger, J. Puc hinger, and R.F. Hartl. A surv ey on dynamic and
sto c hastic v ehicle routing problems. International Journal of Pr o duction
R ese ar ch , 2015.
[37] M. Sc hneider, A. Stenger, and D. Go ek e. The electric v ehicle-routing
problem with time windo ws and rec harging stations. T r ansp ortation
Scienc e , 48:500520, 2014.
[38] A. Sc hrijv er. On the history of the transp ortation and maxim um o w
problems. Mathematic al Pr o gr amming , 91:437445, 2002.
[39] D. Simc hi-Levi, P . Kaminsky , and E. Simc hi-Levi. Managing the supply
chain: the denitive guide for the business pr ofessional . McGra w-Hill
Companies, 2004.
[40] T. Stank, C. Autry , P . Daughert y , and D. Closs. Reimagining the 10
megatrends that will rev olutionize supply c hain logistics. T r ansp ortation
Journal , 54:732, 2015.
[41] P . T oth and D. Vigo. V ehicle r outing: pr oblems, metho ds, and applic a-
tions , v olume 18. Siam, 2014.
[42] M.A. W aller and S.E. F a w cett. Data science, predictiv e analytics, and
big data: a rev olution that will transform supply c hain design and man-
agemen t. Journal of Business L o gistics , 34:7784, 2013.
[43] J. Y ang and H. Sun. Battery sw ap station lo cation-routing problem
with capacitated electric v ehicles. Computers & Op er ations R ese ar ch ,
55:217232, 2015.
19
View publication statsView publication stats