Fleet and revenue management in car rental companies: A literature review and an integrated conceptual framework

TLDR: In this article, the authors present, define and structure the car rental fleet management problem, which includes operational fleet management issues and problems traditionally studied under the revenue management framework, and identify research directions for the future development of the field.

Abstract: This paper aims to present, define and structure the car rental fleet management problem, which includes operational fleet management issues and problems traditionally studied under the revenue management framework. The car rental business has challenging and distinctive characteristics, which are mainly related with fleet and decision-making flexibility, and that render this problem relevant for academic research and practical applications. Three main contributions are presented: an in-depth literature review and discussion on car rental fleet and revenue management issues, a novel integrating conceptual framework for this problem, and the identification of research directions for the future development of the field.

Figures

Figure 1: Literature review structure. This gure represents the structure of this section; the blocks represent the type of papers reviewed, which are divided in three main categories. The last category is divided in smaller sub-categories, by sub-problem.

Figure 3: Heat map of the car rental eet management literature. Based on the proposed framework, this heat map aims to show where most research has been focused and what areas are still developing. The grey intensity scale illustrates the number of papers focused on each problem and the integration of problems: from light grey (1 paper) to dark grey (8 papers)

Table 2: Methods and approaches

Figure 4: Research directions referenced to the proposed framework

Table 1: Factors that in uence the size of instances tested

Figure 2: Framework for the car rental eet management problem. The blocks represent di erent subproblems and the links between them the existing conceptual connections. Other relevant inputs are also represented.

Full text

1
Fleet and revenue management in car rental
companies: a literature review and an
integrated conceptual framework
Beatriz Brito Oliveira
∗ †
, Maria Antónia Carravilla
∗
, José Fernando Oliveira
∗
∗
INESC TEC and Faculty of Engineering, University of Porto
†
Corresponding author: beatriz.oliveira@fe.up.pt
Abstract
This paper aims to present, dene and structure the car rental eet management problem, which
includes operational eet management issues and problems traditionally studied under the revenue
management framework. The car rental business has challenging and distinctive characteristics, which
are mainly related with eet and decision-making exibility, and that render this problem relevant for
academic research and practical applications. Three main contributions are presented: an in-depth
literature review and discussion on car rental eet and revenue management issues, a novel integrating
conceptual framework for this problem, and the identication of research directions for the future
development of the eld.
Keywords:
car rental; eet management; revenue management.
1 Introduction
This paper aims to present, dene and structure the eet management research focused on the problems
faced by car rental companies. The focus on the car rental context arises from the interesting and chal-
lenging idiosyncrasies of its eet and decision-making processes, which have some structural dierences
when compared to other transportation sectors more traditionally studied in the literature. Moreover,
car rental is a growing business, comprising $ 27.11bn in revenue in 2015 in the U.S.  which represented
a 4% improvement over the previous year  while the average car rental eet grew 5% (Auto Rental
News, 2015). This growth trajectory has been steady since 2010 and is forecasted to continue. From
2016 to 2021, the global car rental industry is expected to grow 5.6% (CAGR), due to increasing tourism
activities, the globalization of operations, and the global rise of income levels (ReportsnReports, 2015).
The car rental eet management problem embeds decisions that are traditionally framed within dier-
ent strategic levels and studied by dierent research areas. The main decisions are related with clustering
locations that will share the same eet, deciding on the eet size and composition, distributing eet
amongst rental stations, deciding on prices, selecting which reservations to accept, and assigning these
reservations to specic vehicles. In a real-world setting, these decisions are not only linked by close inter-
actions but also by overlapping decision-making time horizons. In fact, one of the main characteristics
of the car rental eet that motivates this study is its inherent exibility. On the one hand, the eet is
signicantly easy to move and re-locate, enabling e.g. the use of strategic eet balancing decisions often
referred to as empty transfers. On the other hand, there is also a exibility on the decision-making
process that often renders the traditional hierarchical overview of eet decisions too rigid. For example,
the acquisition and removal of cars to and from the eet is signicantly exible, as these contracts are
often incentivized with small lead times, and frequently throughout the year.
Due to the relatively small number of papers that deal with this problem so far, this review aims to
be exhaustive within its scope, which comprises quantitative methods that were developed to support
decisions related with car rental eet management. It is structured in three main parts. Firstly, the
seminal works that launched the interest in the eld are reviewed, which are generally accounts of the
early implementation of eet and revenue management systems in car rental companies. The second
part is devoted to the main works, which have structured the eld and set the ground for future works.
These will be the focus of the third part, which is divided in smaller sections related with the type
of decisions: the clustering of rental locations in groups that share the same eet (pools) and the eet
management within each pool, which comprises operational decisions, revenue management decisions and
the integration of both.
http://dx.doi.org/10.1016/j.omega.2016.08.011

2
Arising from the literature review, a conceptual framework is proposed to structure the car rental
eet management problem. The literature in the area is scarce and somewhat concentrated in only a
few of the problems; however, the interest in this eld has been growing in the past years and expanding
to dierent sub-problems within this scope. The framework herein proposed aims to contextualize the
relations between the dierent sub-problems, and is motivated by the need to support the development
of methodologies that are applicable in real-world settings.
One of the main contributions of this work is also the proposal of four research directions. These
are based on the framework and literature review and are related with the increase of the realism and
applicability of the existing methods and the exploration of dierent levels of integration of the sub-
problems.
The remainder of this paper is thus structured as follows. In Section 2, the problem is introduced
and described in detail; moreover, some lessons learned from research in other transportation sectors
are discussed. Then, in Section 3, the literature review on the car rental eld is presented and discussed.
The framework for the problem is then presented in Section 4 and, based on it, the research gaps are
identied. Four main research directions for the future are also proposed in this section. Finally, some
conclusions are drawn and the main contributions and limitations of this work are discussed in Section 5.
2 Problem and contextualization
In this section, the problem of eet management in the car rental industry, which incorporates several
interconnected sub-problems, is presented. The goal of this section is to informally describe and contex-
tualize the main business decisions, with no specic intention to structure and thus limit the problem
denition.
Fleet management is indeed a mature topic of research in other transportation elds. From some of
these elds, such as the airline industry and maritime or rail-freight transportation, parallels can be drawn
with the car rental business and thus useful lessons can be learned. Nevertheless, there are structural
dierences that support the need for a more specic treatment of the car rental business and these will
also be presented.
2.1 Fleet management in the car rental business
The goal of this section is to broadly present the car rental eet management problem and its main
decisions. In fact, the car rental business protability is heavily dependent on its eet and all decisions
that concern it. These eet decisions span across all strategic levels of the company, from the network
design decisions to specic-vehicle maintenance requirements. The following description focuses on the
main decisions dealt with within this scope regarding the network design, the denition and utilization
of the eet, and the management of booking requests and consequent schedules for each vehicle.
Network
In bigger car rental companies, the rental stations are usually aggregated in pools  groups of
stations that share the same eet. These pools are independent from other administrative divisions (e.g.
regional divisions) although they can overlap; therefore, there is a certain exibility to change and adjust
them. In fact, this is not a one-time decision by nature; the pool design may be frequently reshued
as a means to e.g. meet seasonal changes in demand patterns across locations. There is a specic set of
cars assigned to each pool, to be shared by the rental stations that form it. Within the pool, the specic
location of the car at a certain time depends on its status: if it is fullling a reservation, idle at a certain
station, or under maintenance at a certain workshop. Ultimately, some cars may even be outside their
pool, if a customer picks-up a car in a rental station that belongs to the pool but returns it to a station
outside the pool.
Fleet denition
A car rental eet is composed of a number of cars of dierent types (rental groups).
These groups may be substitutable, which will be discussed later when the reservation proceedings are
described, thus connecting the decisions of how many cars of a certain group to have? for the dierent
rental groups. Overall and generically, the size of the eet is mostly determined by the company's
strategic positioning and available investment, which is a one-time decision and out of the scope of the
eet management problem studied. Nevertheless, the operational adjustments made to increase and
decrease the eet, usually within a pool, are critical for a proper eet management and will be herein
analysed.

3
A eet management problem deeply linked, or even included, in deciding the size adjustments, is
related with the vehicles to acquire and remove from the eet. Actually, buying and selling vehicles can
play a very relevant part on the company's protability. This part of the process is extremely dependent
on the type of car rental company. Some car rental companies are part of vertically integrated business
groups, and thus have a close access to a manufacturer and/or to a wholesale reseller. For these companies,
acquiring vehicles can be compared to a leasing contract, where a specic service deadline is dened for
each car; they can also have access to discounted prices or other amenities. As for removing the cars
from the eet, they are more protected against uctuations in the used car market, for example, as the
responsibility to dispose of the stock falls on the reseller company of the group, or at least is shared
with it. For the remainder of the companies, however,
how
the vehicles are acquired and removed from
the eet is as important as
when
. These decisions can also be signicantly exible, yet this depends
on the mode of acquisition/removal. In fact, although some acquisition contracts must be made with
some antecedence, the assignment of new cars to the respective pools can be made with a short notice, if
needed. The decisions on the removal of cars from the eet are also extremely important, especially since
they can and should be made vehicle-by-vehicle. If the company wants to sell back the used car, specic
information, such as the odometer values, is critical to decide on sell dates (Lacetera et al., 2011).
Fleet utilization
Another critical decision is how to divide the existing eet among the rental stations,
within a pool. This is a critical aspect since the majority of the operational costs in car rental are related
with idle eet. That is to say, the ideal operational goal of car rental companies would be to have 100%
of the eet occupied 100% of the time. These decisions are extremely exible, as the eet levels at each
station are constantly being changed due to incoming returns and pick-ups. Moreover, due to imbalances
on demand and the possibility to
rent here, return there
, there may exist the need to empty reposition
the vehicles between stations, either with a driver repositioning a specic car or by transferring a batch
of vehicles by truck. These transfers, which are critical for balancing the eet levels across the pool, or
among pools, are extremely costly. These costs may be reduced with proper planning methods.
Booking requests and vehicle scheduling
Fleet management in car rental companies also includes
the task of assigning specic vehicles to booking requests. These requests can be made with some
antecedence (reservations), enabling a pre-plan of this assignment, or by walk-in customers, which require
a vehicle on the y from a specic rental station. In some companies, this assignment is decided by the
rental station sta. However, for other companies, this is planned in a somewhat centralized level (e.g.
pool level), especially for rental groups which have a smaller number of cars available (e.g. luxury cars).
Furthermore, it is also important to schedule the planned maintenance for the eet vehicles. This can be
done simultaneously with the scheduling of the reservations.
The main characteristics of a booking request are: the desired renting group, the pick-up (or check-
out) date and station, and the return (or check-in) date and station. If there are no cars from the desired
group available, it is common practice to oer an
upgrade
, i.e. a vehicle from a better rental group for
the price of the originally requested group. Also, as a last resource to avoid a lost sale, some companies
oer a possibility to
downgrade
, i.e. get a vehicle from a worse rental group for a discounted price. It
is because of these strategies that it is critical for car rental companies to manage their eet integrating
all rental groups.
Due to the close links between demand and eet occupation, and the importance that occupation
has on the operational eciency and cost structure of the company, the decision to accept or reject
booking requests is also important when managing the eet. Although some companies may full all
booking requests in order of arrival as long as there is capacity available, other companies manage
demand by saving capacity for more protable reservations that may arrive later, either by using complex
segmentation, capacity allocation or pricing methods or simply by heuristically prioritizing reservations.
Uncertainty
The eet management problem in car rental is severely aected by uncertainty. Demand
uncertainty is the issue most recognized and addressed by car rental companies, by in investing in accurate
forecast methods, for example. Nevertheless, other factors bring uncertainty to the problem, with signif-
icant impact on eet management. For example, when acquiring new vehicles and removing old vehicles
from the eet, the costs and prots associated with these decisions have a degree of uncertainty that can
have a signicant impact on the nal decisions. Also, the availability of the eet is often inuenced by
uncertainties such as unplanned vehicle maintenance and repairs or delayed car returns.

4
2.2 Lessons learned from other transportation sectors
A myriad of sectors and industries have been using quantitative approaches to optimize or improve their
eet management processes. From maritime transportation to humanitarian aid, the need to eciently
manage a eet of vehicles is extended across strategic levels and business functions. Academic research is
more prominent in certain elds, namely in those where transportation is the core business; nevertheless,
interesting and innovative applications have been rising. The process of retrieving the lessons learned
from other sectors is often hindered by the lack of standardized problem names. For example, `assignment'
is a concept with several interpretations in terms of scope, inputs and outputs. `Assignment' can either
refer to allocating parts of the eet to a specic location, to assigning a specic vehicle to a specic
order/demand, or to assigning a type of vehicle to a type of order/demand. Despite these diculties,
in this section there is an attempt to understand what type of research has been developed in the eet
management context in dierent sectors, what parallels can be found with the car rental business, and
what dierences and challenges hinder its straightforward application on this eld.
Airline industry
The airline industry is a traditional comparison sector in the car rental literature.
In terms of operational eet decisions, there is indeed a strong body of research in this eld regarding
eet, tail and crew assignment, amongst other issues. For a thorough review on assignment in the airline
industry see Sherali et al. (2006). Moreover, as a pioneer in the eld of revenue management, the airline
industry is also often referred when dealing with capacity control decisions in car rental.
The focus on eet management problems in the airline sector has been mainly motivated by practical
issues the airlines have been faced with. For example, Salazar-González (2014) describe the analytical
approaches used in a real application: a research project developed alongside a regional carrier, which
worked on issues such as eet assignment, as well as aicraft routing and other problems. These approaches
were validated for real-world instances and are currently being employed by the company.
Fleet assignment in airline relates to assigning aircraft types with dierent capacities and dierent
characteristics to the previously scheduled ights, based on availabilities, operational costs and potential
revenues. This is dierent from the assignment of a specic physical aircraft to a ight, which is called
tail assignment. Flight scheduling, on the other hand, is related with the ight network specication,
including departure and arrival locations and times, working as an input to the assignment problem.
Other critical eet management issue in this eld is the rotation problem, in which an individual aircraft
can be assigned specic routes among those prescribed for its own type, while satisfying maintenance
constraints. (Sherali et al., 2006; Clausen et al., 2010)
The tail assignment problem has strong similarities with the vehicle-reservation assignment problem
in car rental, as the services for each specic physical vehicles are being scheduled. Also, the allocation
of vehicle types to ight legs can be comparable with tactical decisions in car rental. In fact, time-space
networks and the representation of connections and ight legs with arcs is often used in this eld; this is
also frequently considered in some car rental eet decisions in Section 3. However, the main dierence that
motivates the specic study of the car rental problem is the mobility of the eet, namely the possibility
to empty reposition the vehicles. This exibility comes from the costs that, although important, are still
low enough to make these decisions protable; in the airline industry, however, the costs to do the same
are prohibitive. Moreover, in the car rental industry, the process of buying and selling vehicles comprises
important decisions that, due to the business characteristics, turn out to be essentially operational,
namely as far as timings and acquisition modes are concerned. In the airline industry, these variables
are not usually considered alongside other eet management issues and the decisions are taken in a more
hierarchical and sequential way, once again due to the heavier costs associated with eet decisions in this
sector.
Maritime transportation
Research on maritime transportation can bring interesting insights to car
rental eet management, especially regarding the decisions on size/mix. In this industry, the decisions
on how many ships of each type are needed to meet demand are made periodically and often multi-stage
approaches are used, thus representing the eet renewal problem. Pantuso et al. (2014) present a thorough
survey on these problems. The authors also state that models on eet size often include the decision on
assigning specic ships to pre-determined routes as well. In this industry, demand uncertainty is high,
such as in car rental. However, the supply is much slower to adapt to peaks in demand as the lead time to
acquire new ships is signicantly higher than the one to acquire new cars. Nevertheless, one of the lessons
learned from this industry should relate with the detail given to the mode of acquisition and disposal of
the vehicles from the eet. As examples, issues often considered are the possibility of chartering in and

5
chartering out, or laying up, i.e. keeping the ship idle at a specic location with reduced crew and costs.
Also in this eld, it is critical to accurately represent eet heterogeneity.
In summary, research developed on maritime transportation eet management may bring signicant
insights for the car rental business, especially when the heterogeneous eet renewal process is considered.
Nevertheless, once again, the exibility in buying and moving car rental eet brings important advantages
to the process and signicant improvements may be gained by considering it explicitly (versus directly
applying the research developed for maritime transportation to car rental).
Rail-freight
In this context, the empty vehicle redistribution is core. Deciding the railcar distribution,
i.e. where to send empty railcars to meet the next order, can be seen as an assignment to specic
customers. Yet this is usually made at bulk, and not specically for each vehicle. This problem has been
studied since the 1990s (Spieckermann and Voÿ, 1995; Sherali and Suharko, 1998); nevertheless, it is still
relevant and motivating research up to this date (Gorman et al., 2011). From these works, many parallels
can be driven with the car rental business, as the mobility and exibility of the vehicles is similar. One
important idiosyncrasy of the car rental business, however, is that orders must be met without delay,
at the risk of being lost, and cannot be, for example, backordered. In rail-freight, there is also plenty
of research on eet size, a problem commonly integrated with empty reposition decisions, such as in
Sayarshad et al. (2010); nevertheless, this sector usually deals with a more homogeneous eet than in car
rental and substitutability between eet types is not such a critical issue.
Trucking
In the truckload carrying industry, several eet management problems arise that have some
parallels with the car rental industry. Powell (1991) reviewed optimization models and algorithms for
problems such as the assignment of drivers to pending loads or the distribution of vehicles among locations
and dynamically moving them to meet new demand. Several constraints make this problem signicantly
dierent from the car rental one, such as maximum tour length restrictions, time windows on pick-ups
and deliveries, and the possibility to seize backhaul opportunities. Nevertheless, research in this area has
been growing and, for example, eet size and balance has been on the focus of recent works (ak et al.,
2011; Carbajal et al., 2012).
Shared-use vehicle services
Shared-use vehicle services, such as car or bike sharing, are not a
tra-
ditional
transportation sector per se, since the academic and practical interest in this topic has emerged
in the past decades. Nevertheless, since the mid-1980s, there is a growing interest in the academic com-
munity in this sector, related with dierent operational problems (Shaheen, 2013) that have also several
similarities with the car rental eet management problem. In fact, these similarities may allow for lessons
from vehicle sharing systems to be learned by the car rental industry and vice versa.
Barth and Shaheen (2002) broadly dene shared-use vehicle systems as eets of vehicles used by
dierent users throughout the day. The authors propose a framework for classifying these systems, which
can follow dierent operational modes, such as car sharing, where a network of strategic parking locations
is available for the user to pick-up and return the car, and station-car, where vehicles are deployed near
metropolitan rail stations to be used by rail commuters. Gavalas et al. (2015) present a comprehensive
review of algorithms for the management of shared-use vehicle systems, enabling a comparison of both
sectors in terms of issues, goals, scope and approaches. The authors review the methods applied regarding
the design of the vehicle sharing system  including network location, eet size and deployment between
stations , which shows several similarities with the car rental business. The authors also focus on
customer incentivisation schemes to help the balancing the eet and on the operational improvement of
the vehicles empty transfers.
In fact, key lessons can be learned from this sector especially regarding the deployment of the eet
between stations in order to meet unbalanced demand. Nevertheless, although the motivation is similar in
both sectors, the business and operational approaches dier signicantly (e.g. pricing schemes and vehicle
reposition strategies). The main dierences between the two sectors are related with the heterogeneity
of the car rental eet, and the scale and scope of the operations. The latter issue arises from the fact
that vehicle-sharing systems have a daily focus, as dened by Barth and Shaheen (2002), and are usually
centered in a single metropolitan area, while the car rental scope is usually wider in terms of time horizon
and geography. These dierences may lead, for example, to a greater need of detail (e.g. pricing in the
car rental industry is usually almost individualized per reservation).
Nevertheless, the dierences between the two sectors are becoming blurry as shared-use vehicle systems
get operationally more complex (e.g. by introducing heterogeneous eets). Therefore, besides the lessons

Frequently asked questions

Q1. What are the contributions mentioned in the paper "Fleet and revenue management in car rental companies: a literature review and an integrated conceptual framework" ?

This paper aims to present, de ne and structure the car rental eet management problem, which includes operational eet management issues and problems traditionally studied under the revenue management framework. The car rental business has challenging and distinctive characteristics, which are mainly related with eet and decision-making exibility, and that render this problem relevant for academic research and practical applications. Three main contributions are presented: an in-depth literature review and discussion on car rental eet and revenue management issues, a novel integrating conceptual framework for this problem, and the identi cation of research directions for the future development of the eld. 

Q2. What are the future works mentioned in the paper "Fleet and revenue management in car rental companies: a literature review and an integrated conceptual framework" ?

In this process, three main contributions may be highlighted due to their relevance not only to the aforementioned goal but especially to support future developments in this eld of research. Finally, based on this framework, four main research directions for the future are discussed. Some interesting future work may therefore also lie in the expansion of this framework and resulting work to these systems. Overall, the authors believe the framework proposed, and its resulting contributions, will assist the eld in its development where the focus on the business characteristics allied with a strong methodological background will allow the application in real-world settings of the research developed. 

Q3. Why is the decision to accept or reject booking requests important?

Due to the close links between demand and eet occupation, and the importance that occupation has on the operational e ciency and cost structure of the company, the decision to accept or reject booking requests is also important when managing the eet. 

Q4. What is the main downside of this approach?

The main downside of this approach is not the capping assumption on the transfer time, which can be guaranteed by the pool design, but resides on the limitation imposed on the transferring schedules. 

Q5. How is the decision on price updates controlled?

The decision on price updates are controlled by an adaptive heuristic procedure, which is based on actual and desired occupation levels. 

Q6. What is the main reason why a company should manage its eet?

From maritime transportation to humanitarian aid, the need to e ciently manage a eet of vehicles is extended across strategic levels and business functions. 

Q7. What are the main drivers of revenue in car rental?

In car rental, it is possible to observe that the main drivers of revenue are the demand, which is highly price-sensitive, and the occupation levels, which have a strong impact on the allocation of operational costs. 

Q8. What is the author's proposal for a relax-and-x heuristic?

The authors propose a relax-and- x heuristic procedure, which includes a constraint based on local branching that enables and controls modi cations between iterations. 

Q9. What is the main function of determining the optimal eet size and mix for each?

The authors consider that determining the optimal eet size and mix for each location on a daily basis is a primary function of revenue management . 

Q10. What are some of the problems wherein considered are often studied under the revenue management framework?

Some of the problems wherein considered are often studied under the revenue management framework, namely the ones where pricing and capacity allocation decisions are concerned. 

Q11. What are the main considerations in the upgrading of a car?

In this work, the authors discuss at length the upgrading strategies and their impact in car rental and state that the two main considerations are fairness and scope. 

Q12. What is the level of detail required to solve the pricing problem?

Pricing is an issue that also requires external inputs such as market information regarding competitor prices, with a signi cant level of detail: usually, per reservation type. 

Q13. What are the constraints that make this problem different from the car rental one?

Several constraints make this problem signi cantly di erent from the car rental one, such as maximum tour length restrictions, time windows on pick-ups and deliveries, and the possibility to seize backhaul opportunities. 

Q14. What are the main motivations for rms to practice yield management?

The authors present the main motivation for rms to practice yield management and present the traditional tools for capacity control (booking limits, protection levels and overbooking) as well as other extensions.