biosystems engineering 105 (2010) 41–50

Available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/issn/15375110

Research Paper

Conceptual model of fleet management in agriculture

CG Sørensen, DD Bochtis*
University of Aarhus, Faculty of Agricultural Sciences, Department of Agricultural Engineering, Blichers Allé, 8830 Tjele, Denmark

article info
Agricultural fleet management is viewed as farmers’ or machine contractors’ decision-
Article history: making concerning, for example, resource allocation, scheduling, routing, and real-time
Received 14 May 2009 monitoring of vehicles and materials. In order to optimise this management task, fleet
Received in revised form management tools are used for decision support to improve scheduling, routing and other
18 August 2009 operational measures for a fleet of agricultural machines. Additionally, fleet management
Accepted 15 September 2009 involves the process of supervising the use and maintenance of machines and the asso-
Published online 25 October 2009 ciated administrative functions including the coordination and dissemination of tasks and
related information for solving the heterogeneous scheduling and routing problems. In
order to analyse the complex situation of how to develop an effective Fleet Management
System, a conceptual model was developed based on a participatory approach and
subsequent system analysis. This involved identification of the scope of the system,
identification of user requirements, conceptual modelling, identification of actors and
decision processes, and information-needs modelling. A holistic view and scope of the
system is presented together with the system constraints. The results build on extracted
information from selected actors in the agricultural domain. The systems components are
depicted and linked to the subsequent conceptual model of the overall system as an outline
for a subsequent tool development and implementation.
ª 2009 IAgrE. Published by Elsevier Ltd. All rights reserved.

1. Introduction in-field productivity and the maximum theoretical produc-

tivity defined by the maximum operating speed and
The structural development and the imposed requirements maximum working width (Witney, 1988). Especially for larger
within agriculture imply that innovative technology and machines, it is important to maintain a high efficiency as the
knowledge management will be decisive for the future arable non-productive time elements represent a greater propor-
farmer (Fountas et al., 2006). Especially, the maximisation of tional loss in potential machine production (Sørensen, 2003;
agricultural machine productivity is an important element in Søgaard and Sørensen, 2004).
the continued efforts of planning and controlling resource A preliminary step in the direction of achieving increasing
input in arable farming. Different forms of driving instruc- operational efficiency is a renewed focus on the usage of
tions and advice on optimising the execution of field opera- advanced Information and Communication Technology (ICT)
tions provide the possibility of influencing a number of factors systems and formal management models in agriculture
pertaining to the operational efficiency (e.g., Palmer et al., (Sørensen et al., 2008). In many ways, this development may
2003; Sørensen and Nielsen, 2005; Bochtis et al., 2007). The be compared with the mechanization of agriculture in the
operational efficiency expresses the ratio between the actual 1950s and 60s. At that time, the introduction of the tractor,

* Corresponding author.
E-mail address: [email protected] (DD Bochtis).
1537-5110/$ – see front matter ª 2009 IAgrE. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.biosystemseng.2009.09.009
42 biosystems engineering 105 (2010) 41–50

Nomenclature ICT Information and Communication Technology

KPI Key Performance Indicator
ERP Enterprise Resource Planning
SSM Soft Systems Methodology
FMS Fleet Management System

combine harvester and other machinery items radically contractors’ in terms of providing the conceptual guidelines
changed the nature of field operations towards more auto- for optimised decision-making concerning resource alloca-
mation, both in terms of technology and management tion, scheduling, routing, real-time monitoring of vehicles and
measures. In the long perspective, research shows that materials and timely field operations or customer orders, for
a paradigm shift from large machines to smaller and more example.
intelligent multi-robot systems, which can establish and
nurse, for example, plants at an individual level is expected
(Blackmore et al., 2005; Fountas et al., 2007). Such a develop- 2. Fleet management concept in agriculture
ment will increase the demand for advanced management
tools, like fleet management tools, for scheduling, monitoring Logistics may be defined as ‘‘the provision of goods and
and on-line coordination of multiple vehicles. services from a supply point to various demand points’’ (Eilon
Fleet Management Systems (FMS) have been available in et al., 1971). Other definitions include ‘‘Logistics is the science
the industrial domain, such as the transport business, for of planning and implementing the acquisition and use of the
a number of years. The first generation of FMS was relatively resources necessary to sustain the operation of a system’’
simple software applications coping with a number of simple (Cooke, 1993). In an industrial context, a complete logistics
functionalities such as vehicle tracking components as a basic system involves transporting raw materials from a number of
function (e.g., Crainic and Laporte, 1998; Mele, 2005). suppliers, delivering them to the factory plant for processing,
Currently, these systems have evolved into complete enter- transporting of the products to different depots and in the end
prise management tools linking together all parts of the distributing them to the customers. In this case, both the
business. The new trend clearly dictates increased manage- supply and distribution transportation require effective
ment sophistication in terms of turning the tools into plan- management procedures in order to optimise routes and
ning tools (Van Heijden and Marchau, 2002; McLoad, 2005). It costs. These procedures are part of the overall Supply Chain
includes real-time asset management focussing on current Management of the company (Adam and Ebert, 1992). Poten-
fleet locations and the prediction of planned tasks. These tial costs savings include reduced transportation and labour
systems today offer a broad range of functionalities including costs.
tight integration with internal Enterprise Resource Planning Logistics within agriculture may be viewed as an inclusive
(ERP) systems and systems located at the customer. Specifi- management notion establishing a comprehensive view of the
cally, an extensive use of real-time data and wireless material flow in the production process (Hahn, 2006). For
communications are seen together with increased intelli- example, in dairy operations, the logistics concerns the loca-
gence for real-time planning, where industry developers tion of the different functional units (stable parts, parlour,
identify those parameters as the primary drivers for the feeding storage, etc.) related to each other and combined with
current development (Maddocks, 2005; Bernard et al., 2008). the management systems employed. A logistical improve-
Agriculture is currently undergoing a significant develop- ment on the farm involves more rational work flows and
ment, where more advanced machinery as well as informa- operations. The basic logistical notion is that all operations
tion technology are being implemented (Pesonen et al., 2008). and actions must create an added value in the process chain
In principle, these technologies enable the implementation of through the process of having the right thing, at the right
the analogous fleet management tools as seen in the industry place, at the right time.
domain. However, the inherent biological and dynamic nature As part of the management of the logistics, fleet manage-
of agricultural operations together with the experience of ment may be used as the practical tool managing a fleet of
lower general user acceptance have proved to inhibit inte- vehicles to improve scheduling, operational efficiency, and
gration of current FMS into the agricultural domain (e.g., Gelb effectiveness (Auernhammer, 2001). Additionally, fleet
and Offer, 2005). management involves the process of supervising the use and
The objective of this paper is to derive the contextual maintenance of vehicles and the associated administrative
requirements and develop a conceptual model for fleet functions including the coordination and dissemination of
management in agriculture. The concept includes the deci- tasks and related information for solving the heterogeneous
sion support and optimisation of operations executed by scheduling and vehicle routing problem.
a fleet of agricultural machines (non-autonomous or autono- In intensive agricultural production systems, large
mous). Also, these operations include the configurations of amounts of capital are invested for the acquisition, operation
teams of identical machines (e.g., multiple forage harvesters), and maintenance of highly-efficient machinery requiring
co-operative machines (e.g., combines and transport units, a high utilization rate. Hence, the execution of field operations
fertilizers and supply units), or machines in cooperation with by one machine or many cooperating machines needs to be
labourers (e.g., orchard harvesting). The hypothesis depicts carefully planned and executed in order to achieve maximum
that the generalised notion of fleet management will be able to operations efficiency. For example, large scale harvesting
meet a number of challenges facing farmers’ or machine involves complex planning efforts, due to the uncertainties
 biosystems engineering 105 (2010) 41–50 43

Fig. 1 – The total supply chain of integrated field operations.

associated with yield, weather, machine performance etc. The both cases, the size and composition of these teams must be
planning of such operations typically involves four stages that carefully selected and dispatched, and the teams must be
are highly interconnected. These are: harvesting, out-of-field coordinated in the field, in order to increase their collective
removal of biomass, rural road transportation and public road work rate.
transportation (Fig. 1). The corresponding machinery system The existing agricultural fleet management tools are
includes: harvesters, transport units, medium and high oriented toward off-line management systems (Sørensen and
capacity transport trucks, and unloading equipment between Thomsen, 2006). This type of management gives the appro-
each pair of successive stages. All these comprise a complex priate fleet size and composition as well as the machine
system in which the individual efficiency of each machine allocation and scheduling for a given operation. The proce-
unit depends on the performance of the system as a whole or dure determines the allocation of the fields (or field parts) to
the supply chain. the available machines of the fleet, the assignment of these
Large scale operations of similar planning complexity to machines to the available supporting vehicles, and the
harvesting or the removal of material from the point of origin number of deposit or refilling units (facility units) that will be
(‘‘material output operations’’) are the ‘‘material input oper- used, described by their type (e.g., mobile, non-mobile), their
ations’’ like spraying and fertilizing. These operation also location, capacity, cycle time, etc. Although this type of
involve multi-machinery systems that potentially include one machinery management relates to an off-line planning
or a number of self-propelled or tractor–trailer units and one system, it constitutes a very complex process and involves the
or more transport units. Furthermore, with the prospect of interactions between the farm machinery system and bio-
wide implementation of precision farming techniques such logical and meteorological subsystems such as crop, soil and
as site-specific material application, real-time planning is weather conditions. A traditional method of evaluating
indispensable. systems with interacting machines or components is cycle
In addition to the mentioned agricultural operations that analysis, involving a method to account for all time spent by
naturally incorporate the fleet machinery concept, the growth each machine in the system and this procedure can be useful
of the range of available self-propelled machines including for identifying suitable transport systems in field operations
fertilizer spreaders, plant protection sprayers, mowers, etc., (Buckmaster, 2006). The disadvantage of such an approach is
motivates the adoption of ‘‘team work’’ for the execution of that the cycle diagram constitutes a static instrument that
the corresponding agricultural operations. This adoption has does not take into account variability in speed, yield transport
the advantages of a collective behaviour and allows scheduled distance and time evolution. Some powerful optimisation
work to be carried out on time. On the other hand, if large methods have been adopted in order to deal efficiently with
teams of smaller autonomous machines are to replace smaller the inherent large number of decision and state variables.
groups of heavier machines in the future, ‘‘fleet management’’ Søgaard and Sørensen (2004) presented an approach involving
will also play a key role in maximizing the overall efficiency. In the development of a non-linear programming optimisation
44 biosystems engineering 105 (2010) 41–50

model based on a level of aggregation consistent with the tracks (for one or for multiple fields), covered by one machine
accessible and existing data related to machinery sets, crops, or by a fleet of machines was presented by Bochtis (2008). This
weather and timeliness of operations. Busato et al. (2007) algorithmic approach improved the field efficiency of the
developed a dynamic discrete-event simulation model in machines by minimizing the total (in-field and out-field) non-
order to optimise the wheat harvesting and transport opera- working travelled distance. An implementation of the
tion accounting for field size and shape, field distance to silo, previous procedure in conventional agricultural machines
yield and resources available. An enhanced approach of this supported by auto-steering systems was presented in Bochtis
system was presented by Berruto and Busato (2008), where the and Vougioukas (2008).
event-oriented simulation is combined with linear program- Fleet management is an integral part of the overall
ming for biomass supply chain evaluation. The developed management system in a purposeful organisation comprising
model considers the interaction among resources and the tolls like ERP, overall information systems, etc. ERP is an
effect of a number of limiting factors on the performance of industry notion for a wide set of management activities which
the whole chain. The tool is suitable for detailed evaluation of support all essential business processes within the enterprise.
the system efficiency by considering multiple parameters The management system supports management activities on
(e.g., field yield, shape, size, transport distance and working all levels as well as providing for the identification of Key
chain composition). Foulds and Wilson (2005) reported on two Performance Indicators (KPIs) (Folinas, 2007). Typically, ERP is
practical harvesting case studies (rape seed and hay harvest- integrated with a database system and will often include
ing) based on empirical data from commercial enterprises and applications for the finance and human resources aspects of
where a scheduling model was developed based on integer a business.
programming as well as heuristics methods. Basnet et al.
(2006) extended the latter approach to include scheduling
operations for multiple fields. The main conclusion from 3. Methodology
these approaches is that the scheduling of harvesting opera-
tions is a significantly different scenario from those repre- 3.1. Participatory approach
sented by the currently available scheduling models, due to
difficulties caused by the interchangeability of the workers As a step in the procurement of requirements on systems for
and equipment employed, the need to schedule each of these integrated fleet management, a goal-directed study involving
resources individually, the presence of time lags and the fact relevant actors and stakeholders was carried out. The meth-
that travel times should be taken into account. Higgins and odological approach involved a participatory approach and
Davies (2005) developed a capacity stochastic planning model analysis, extracting current operations management chal-
for sugarcane transport that estimates the number of loco- lenges facing farmers in terms of increased productivity
motives, shifts, and bins required as well as the delays in the demands as well as increased compliance requirements from
harvesting operations resulting from harvesters waiting for society. The focus was on extracting the different specific
bin deliveries. The simulation model was also extended to requirements on fleet management including:
optimise harvester start times using a threshold-accepting
heuristic method.  extracting individual requirements on fleet management in
Although scientific research has contributed to the devel- relation to current and future activities within this area
opment of models for off-line scheduling of field operations  identifying necessary system components as regards soft-
involving fleets of agricultural machines, the actual execution ware and hardware
of these operations is in most cases performed by relying on  identifying the need for supplemental development
the operators’ experience and without any optimal on-line
planning. Some promising approaches regarding the on-line The study was carried out as an interview survey involving
planning have been reported recently. Bochtis and Vougiou- three Danish companies (The Danish Association of Machine
kas, (2007) presented an algorithmic procedure based on Contractors, Danish Agricultural Advisory Service, and Data-
dynamic programming for planning harvesting operations for Logisk Inc) and an Austrian company, PROGIS Software
a fleet of harvesters supported by a fleet of transport carts. The GmbH, which all work with development and maintenance of
optimisation criterion for the planning was a minimization of planning tools for agriculture. The surveys included targeted
the total travelled length of the paths followed by the carts. questionnaires administered by an experienced researcher
Also, the optimisation included penalty factors for the cases during each interview. The questionnaires included both
where a harvester stops its operation while waiting for the closed and open-ended questions and followed established
transport cart. Another implementation of a well-developed guidelines for surveys (Fink and Kosecoff, 1998). Efforts were
method from other scientific areas is the one of the Petri nets. made to avoid any bias in the process of interviewing farmers
Guan et al. (2008) introduced Hybrid Petri nets into the model- by introducing standardised lists of options to be answered. A
ling of farm work flow in agricultural production. This non-random sampling method was considered feasible in
procedure is based on the fact that agricultural operations view of the main objective being a pilot study engaged in an
involve both continuous and discrete events. The simulated exploratory analysis (Lohr, 1999). A high degree of interaction
results revealed that the Hybrid Petri nets model is promising between the interviewer and company was pursued and
for exactly describing the farming process and reallocating likewise, a wide study of the area in question was pursued.
resources in the presence of uncertainties. An algorithmic The themes and checklist included in the interview are given
approach for computing traversal sequences for parallel field in Table 1.
 biosystems engineering 105 (2010) 41–50 45

which the system has meaning and the transformation

Table 1 – Interview components
depicts functionality at system level. Customers are the ones
Theme Checklist influenced by the transformation in terms of benefitting and
Operational planning/ Route guidance suffering from it, the Actors are the entities that carry out the
route planning Messages to and from the driver system activity and the Ownership belongs to the entities with
Material control and management the power to initiate or terminate system activities. Finally,
On-line positioning of vehicles the Environmental constraints represent elements that are
Expediting of work orders
external to the system and imposed on the system. Checkland
Global planning for multiple machines
and Scholes (1999) argue that the CATWOE transformation is
Documentation/ Traceability measures more elaborated since it includes additional and related
traceability Invoicing
elements and, when included, will lead to enriched root defi-
Technical data on driving economy
nitions and hence, better models. Experience shows that
and maintenance
Follow-up on driving times and resting omitting any of the elements in the CATWOE definition will
times cause the analysis to suffer.
Machine monitoring The derivation of a conceptual model is the first step, indi-
Software/hardware Global Positioning System (GPS)
cating the concept of the management system to be designed.
navigation Although the analysis and modelling process can be elaborated
Communication solution in rather clear points, the outcome of each point is actually
Preventive maintenance continuously changing as more knowledge is generated as
Tracking of vehicles a consequence of the improved understanding of the system.
Data acquisition
Systems analysis has a number of stages (Fig. 1). The
process starts with a description of the current system and
situation, continues with the current logical system, the
3.2. Conceptual modelling proposed conceptual model integrating future demands and
finally with the proposed use-cases or functionalities. For the
In order to analyse the complex and soft-systems situations of first stage, the description of the current systems and situa-
how to develop an effective FMS, the Soft Systems Method- tion, SSM was used to analyse the complex environment
ology (SSM) was used (Wilson, 2001). This approach has been (Checkland, 1988). During the design phase of the conceptual
successfully used in many fields of applications (e.g., model, a logical system was constructed for the proposed FMS
Macadam et al., 1990; Kasimin and Yusoff, 1996) and involves and later the functionalities were derived, which can used to
identification of the scope of the system, identification of user develop the functional prototype.
requirements, conceptual modelling, and information needs
as a preliminary step before any detailed direct system
development. The boundaries and scope of a system can be 4. Results and discussion
described in terms of users and use-cases, where users are
entities interfacing with the system (e.g., managers, software, 4.1. Derived expected functionalities of the fleet
databases) and the use-cases describe the functionality of the management
system or what the users want the system to do.
An important step involves the analysis of how the users The requirements put forward by the machine contractors
understand the current situation of managing mobile clearly embrace a fleet management solution, which can fulfil
machinery items in agriculture. The managers of the pilot the information need for transport administration and
farms were asked to give their opinion which then formed the control. The focus is on transport control, route guidance in
basis for developing an initial description of the soft and hard connection with visiting customers, invoicing, data acquisi-
facts and the conflicts and problems that users are currently tion as regards technical and operational data concerning the
faced with, both from an external and an internal point of individual machine, etc. The voiced requirements from
view. The analysis of the current situation forms the basic farmers concentrate more on on-farm functionalities such as
definition of the system and is also the basis for proposing on-line monitoring, route guidance, operations scheduling,
changes to form the conceptual future FMS. etc. Table 2 lists the voiced requirements.
The conceptual model of the FMS of the system was built Derived competences for model development efforts
on the premises embedded in the definition of the system include route planning, analysis and decomposition of
(Checkland and Scholes, 1999). This definition is called the machine operations and various resource optimisation and
root definition and this concept plays a central role in the decision support. The extracted functionalities form the basis
analyses and modelling as it defines the goal of the system for stating the purpose and inherent assumptions of the
and brings forth various perspectives on a system and the proposed FMS.
inherent assumptions. The root definition is devised in the
form of the elements of CATWOE, a mnemonic concept rep- 4.2. CATWOE
resenting the terms Customers (C), Actors (A), Transformation
process (T), Weltanschauung (W), Ownership (O) and Envi- Based on the derived functionalities, the precise definition of
ronmental constraints (E). The core elements are the T and the the FMS was established within the context of the CATWOE
W, where the weltanschauung depicts the world view for definition.
46 biosystems engineering 105 (2010) 41–50

Table 2 – Voiced requirements for the FMS

Data acquisition and control - On-line positioning of vehicles
- Machine monitoring/tracking
- Improved general knowledge of the production process and management
- More effective invoicing and documentation system
- Detailed work time specification
- On-line information search
- Easy and quick access to information for employees
- Data exchange interfaces for existing tractor and implement computers

Predictive planning and - Expediting of work orders

optimisation - Environmental benefits (e.g., soil compaction, resource usage)
- Resource minimisations (e.g., labour, fuel)
- Preventive maintenance (e.g., alarm settings for engine conditions)
- Coordination of multiple machines (farmers, contractors)
- Contextual awareness (automatic acquisition of data dependent on the vehicle situation)
- Route and path guidance
- Benchmarking

Software/hardware/technology - Planning agent

components - Dedicated user-interface
- GPS equipment
- Implementation of research results on coordination and optimisation of the cooperation between
manned and unmanned machines within the portfolio of existing company software solutions
- Implementation of new exchange formats (like eXtensible Markup Language (XML) format) in the
existing company software solutions, with the intent of improve the data exchange between different IT-
systems
- Wireless communication
- Communication with internal/external databases
- Sensor system for continuous status information

4.2.1. Customers Based on the identified elements of the CATWOE, an

The primary customer of the proposed FMS is the farm inclusive root definition is derived. The root definition of the
manager or machine contractor manager and the manage- purposeful activity handled here is: ‘‘a FMS (owned and
ment system as the demanders of data for production and operated on farm/contractor level) to support real-time
operations management. management decision-making of mobile units, by means of
automated acquiring and contextualising of operations data
4.2.2. Actors and external parameters to form a foundation for decision-
The actor is the one operating the FMS, which in this case is making in order to improve the quality of decision-making
the farm/contractor manager or other farm staff. and reduce the time efforts’’.
The next step requires the FMS to be viewed in terms of
4.2.3. Transformation process how the proposed FMS should perform its functionalities. It
The transformation process involves the transformation of involves specifying the requirements which must be met in
operational field data into a form, which can provide the order to accommodate the highly complex and dynamic
foundation for decision-making in crop production, specifi- environment of the agricultural domain.
cally the management of mobile units.
4.3. Fleet Management System requirements
4.2.4. Weltanschauung
The weltanschauung is the hypothesis that drives the FMS As defined in the previous section, a FMS serves several
development. In this case, the view is that operational data is purposes. A dedicated and well-designed FMS must be able to
easily acquired and can be use to improve the management of collect, process and store the required data and be capable of
mobile units. producing aggregate data and documentation as well as
providing information planning and control purposes. It must
4.2.5. Ownership be integrated into enterprise financial and management
The farm/contractor manager is the owner of the FMS in systems as a way to reduce duplication of data-entries and
terms of every day decision-maker responsibility and as the producing data for reporting requirements, both for internal
decider on whether the system is of use or not. and external purposes.
In a general approach to attain operations efficiency in an
4.2.6. Environmental constraints agricultural FMS, the following requirements must be fulfilled:
The constraints influencing the usability and performance of
the FMS includes the expectations of the managers, the 4.3.1. Robustness
required data quality, the reliability and used information The overall scheduled operation must be completed in an
technology (communication devices, server, databases, etc.). optimised way given the available resources due to
 biosystems engineering 105 (2010) 41–50 47

unexpected events such as functional failure or the improper machinery fleet size and composition, the machine alloca-
function of some systems components. tions and scheduling of the given operation. It also allocates
the fields (or field parts) to the currently available primary
4.3.2. Balance optimality vs. velocity machines (e.g., combines) of the fleet, it assigns these
Due to the extensive dynamic environment inherent in agri- machines to the available supporting units (e.g., transport
cultural operations, it is necessary to be able to optimise the carts to combines) and to the number of the facility units (e.g.,
response of the system to these conditions. On the other hand, deposit or refilling units for the cases of harvesting and
there is a trade-off as to the capability to respond quickly to spraying/fertilizing, respectively) that will be used taking into
the dynamic conditions in order to ensure a successful task account their type (e.g., mobile, non-mobile), their localiza-
execution. tion, capacity, cycle time etc.

4.3.3. Extensibility–flexibility 4.4.2. On-line management

The system must have the ability to easily add and remove Fig. 2 shows the processes in the conceptual model as derived
functionalities as well as to reconfigure its architecture in from the definition of the system. The current situation and
order to adapt to the needs of the different requirements of the derivation of the logical system are based on user input,
various applications. This is critical notion in the agricultural while the derivation of the conceptual model and the func-
application domain. tionalities is based on the proposed changes to the current
situation. In order to incorporate the dynamic nature of the
4.3.4. Allocation field operation, and the inherent uncertainties of many of its
Task allocation is an important factor in the architectural parameters (e.g., unknown yield distribution), the adoption of
design. Tasks should be generated and allocated in a way that a closed loop control system, which results in a sequence of
adopts the principle of incremental progression where at any planning, execution and re-planning, is suggested. As indi-
time the current planning is based on the best information cated, plan generation and execution must be linked in
available. a system monitoring effects of actions, unexpected events and
any new information that can attribute to a validation,
4.3.5. Heterogeneity a refinement, or a reconsideration of the plan. An important
The system’s architecture should be able to accommodate any aspect is that supplementary knowledge from observations,
machinery fleet regardless of the specification of the indi- databases, sensors, etc. can be incorporated in order to revise
vidual unit. plans. The closed loop approach makes feasible the imple-
mentation of an on-line decision support system for the
4.3.6. Changeability coordination of mobile machinery units operating in a field or
The system’s architecture should be able to accommodate the in a number of geographically dispersed fields (Fig. 3).
inclusion or the exclusion of a specific mobile unit. This
specific ability is considerably useful due to the short time 4.4.3. Centralized and de-centralized management
window associated with each type of field operation (e.g., Centralized management systems provide the key advantage
harvesting, seedbed preparation, seeding, fertilizing), where of globally optimal plans due to the fact that the decision-
the available machinery resources must be allocated to maker (human or automated system) can take into account all
concurrent field operations. the relevant information conveyed by the members of the
team. However, centralized approaches often involve intrinsic
4.3.7. Adaptation difficulties such as intractable solutions for large machinery
The system must be able to adapt to specific applications by teams due to the complexity of the required algorithms for the
easily adjusting and tuning the relevant parameters.

4.4. Conceptual models main elements

User
User
4.4.1. Off-line management
The off-line planning for agricultural fleet operations includes
three layers of activities: search of operations history, rules
identification, and initial planning. The first layer regards the Current
History Current Logical system
historical data management on critical factors such as the History situation Logical system
situation
labour and machinery input and the biological and meteoro-
logical conditions adhering to the operation in question. The
second layer consists of the identification of the thresholds
and events that are important to monitor, note and/or will
activate automatic behaviours. Based on the identified rules, Use cases Conceptual
the evaluation of the historical data and the available Functionalities Conceptual
model
recourses, an initial planning has to take place before the Functionalities model
execution of the actual operation concerning required opera-
tions, operations urgency, formulation of jobs, operations Fig. 2 – Principle of system analysis and conceptual
specification, etc. Initial planning provides the appropriate modelling (adapted from Checkland, 1988).
48 biosystems engineering 105 (2010) 41–50

Fig. 3 – Derived conceptual model involving a closed loop approach of the on-line decision support system.

global planning systems. Also, the requirements in terms of planning for main units (e.g., harvesters, fertilizers, seeders),
extensive machine-to-machine and centre-to-machine path planning for in-field service units (e.g., material transport
communication often mean that the real-time response is not carts) and routing for inter-filed transport units (e.g., transport
feasible. On the other hand, a de-centralized management carts) (see Fig. 4) should be placed on-board in order to plan
approach to agricultural fleet management architecture using both the a priori information provided by the centralized
provides the advantages of a fast response to dynamic Geographic Information System (GIS) system as well as the
conditions and decreased communication requirements. In updated information of the local sensing measures.
this way, an improved adaptation of the machines to the
changing operational conditions is achieved, since this
adaptation is carried out by locally sensing and responding to 5. Conclusions
the environment.
Algorithms for scheduling, task allocation, machinery This research has shown the potential benefit of using
assignment, area coverage and route and path planning dedicated system analysis methodologies as an initial step to
should be distributed efficiently in terms of the balance the actual design and conceptualisation of a dedicated FMS
between communication and computational requirements. for agriculture. The SSM has been used to target organisa-
For example, dynamic planning tools for area coverage tional business and process modelling through
 biosystems engineering 105 (2010) 41–50 49

Fig. 4 – De-centralized vs. centralized planning distribution.

a participatory approach involving users and stakeholders as optimality and decision process time, a balance between
providers of expected requirements for a proposed FMS that extensibility and flexibility, efficient allocation measures,
will fulfil the user requirements of tomorrow. The proposed heterogeneity, changeability, and adaptation.
system involves a FMS (owned and operated on farm/ The central elements of the conceptual fleet management
contractor level) to support real-time management decision- model involve an elaborated division between centralized and
making of mobile units, by means of automated acquiring de-centralized management measures. Planning tools for
and contextualising of operations data and external param- scheduling, task allocation, machinery assignment, area
eters to form a foundation for decision-making, in order to coverage and route and path planning, should be distributed
improve the quality of decision-making and reduce the time efficiently in terms of the balance between communication
efforts. and computational requirements utilizing both the a priori
Selected derived functionalities of the conceptual fleet information provided by the centralized GIS system as well as
management model include on-line positioning of vehicles, the updated information of the local sensing measures.
machine monitoring/tracking, improved general knowledge of
the production process and management, automatic invoicing
and documentation system, detailed work time specification
references
measures, automatic expediting of work orders, resource
minimisations (e.g., labour, fuel), coordination of multiple
Adam E E; Ebert R J (1992). Production and Operations
machines (farmers, contractors), route and path guidance, etc.
Management: Concepts, Models and Behaviour. Prentice Hall,
The complement of proposed functionalities is associated with
Englewood Cliffs, NJ 07632, 729 pp. USA.
a number of derived requirements securing an efficient oper- Auernhammer H (2001). Precision farming-the environmental
ational performance of the proposed FMS. These requirements challenge. Computers and Electronics in Agriculture, 30(1–3),
and capabilities include robustness, a balance between 31–43.
50 biosystems engineering 105 (2010) 41–50

Basnet C B; Foulds L R; Wilson J M (2006). Scheduling contractors’ auspices of and supported by the European Federation for
farm-to-farm crop harvesting operations. International Information Technologies in Agriculture, Food and the
Transactions in Operational Research, 13(1), 1–15. Environment (EFITA) and the Samuel Neaman Institute for
Bernard K S; Cheung K L; Choy C L L; Wenzhong S; Jian T (June Advanced Studies in Science and Technology.
2008). Dynamic routing model and solution methods for fleet Guan S; Nakamura M; Shikanai T; Okazaki T (July 2008). Hybrid
management with mobile technologies. International Journal Petri nets modeling for farm work flow. Computers and
of Production Economics, 113(2), 694–705. Electronics in Agriculture, 62(2), 149–158.
Bochtis D D; Vougioukas S G (2007). Agricultural Machine Hahn J (2006). Logistics. In: Yearbook Agricultural Engineering
Allocation based on Simulation. In Proceedings of the Second 2006 (Harms H; Meier F eds). Landwirstschaftsverlag, Münster.
IFAC International Conference on Modeling and Design of Higgins A; Davies I (2005). A simulation model for capacity
Control Systems in Agriculture, pp. 147–152. Osijek, Croatia. planning in sugarcane transport. Computer and Electronics in
Bochtis D (2008). Planning and control of a fleet of agricultural Agriculture, 47, 85–102.
machines for optimal management of field operations. Ph.D. Kasimin H; Yusoff M (1996). The use of a soft systems approach in
thesis, A.U.Th., School of Agriculture, Agriculture Engineering developing information systems for development planning:
Laboratory. an exploration in regional planning. Computers, Environment
Bochtis D D; Vougioukas S G (2008). Minimising the non-working and Urban Systems, 20, 165–180.
distance travelled by machines operating in a headland field Lohr S (1999). Sampling: Design and Analysis. Duxbury Press,
pattern. Biosystems Engineering, 101(1), 1–12. USA.
Buckmaster D R (2006). Systems approach to forage harvest Macadam R; Britton I; Russell D; Potts W; Baillie B; Shaw A
operations. 2006 ASAE Annual Meeting, Paper no. 061087. (1990). The use of soft systems methodology to improve the
Berruto R; Busato P (2008). System approach to biomass harvest adoption by Australian cotton growers of the siratac
operations: simulation modeling and linear programming for computer-based crop management system. Agricultural
logistic design. ASABE Annual International Meeting, Rhode Systems, 34, 1–14.
Island, Paper no. 084565. McLoad T (August 1, 2005). Fleet management systems: the future
Busato P; Berruto R; Saunders C (2007). Modeling of grain is here. Fleet Owner.
harvesting: interaction between working pattern and field bin Mele J (August 1, 2005). Fleet management systems: the future is
locations. Agricultural Engineering International: The CIGR here. Fleet Owner.
Ejournal, IX. Manuscript CIOSTA 07 001. Maddocks B (August 1, 2005). Fleet management systems: the
Blackmore B S; Stout W; Wang M; Runov B (2005). Robotic future is here. Fleet Owner.
agriculture–the future of agricultural mechanisation? In: Fifth Palmer R J; Wild D; Runtz K (2003). Improving the efficiency of
European Conference on Precision Agriculture (Stafford J. V.), field operations. Biosystems Engineering, 84(3), 283–288.
pp. 621–628 Wageningen Academic Publishers, The (red) Pesonen L; Koskinen H; Rydberg A (2008). InfoXT-User-
Netherlands. centric Mobile Information Management in Automated Plant
Checkland P B (1988). Informationsystems and systems thinking: Pproduction. Recommendations and guidelines for a novel,
time to unite. International Journal of Information intelligent, integrated information and decision support
Management, 8(4), 239–248. framework for planning and control of mobile working units.
Checkland P; Scholes J (1999). Soft Systems Methodology in Nordic Innovation Centre.
Action. John Wiley & Sons Ltd, Chichester. Søgaard H T; Sørensen C G (2004). A model for optimal selection
Cooke W (1993). The Government Computer Magazine of machinery sizes within the farm machinery system.
‘‘Integrated Logistics’’ December 1993. Biosystems Engineering, 89(1), 13–28.
Crainic T G; Laporte G (1998). Fleet Management and Logistics. Sørensen C G (2003). A model of field machinery capability and
Kluwer Academic Publishers, Boston, Norwell, MA. logistics: the case of manure application. Agricultural
Eilon S; Watson-Grandy C; Christofides N (1971). Distribution Engineering International: The CIGR Journal of AE Scientific
Management: Mathematical Modeling and Practical Analaysis. Research and Development. ISSN: 1682-1130, 5. ISSN: 1682-
Hafner, New-York. 1130.
Fink A; Kosecoff J (1998). How to Conduct Surveys: a Step-by-Step Sørensen C G; Nielsen V (2005). Operational analyses and model
Guide. Sage Publications, Thousand Oaks, USA. comparison of machinery systems for reduced tillage.
Folinas D (2007). A conceptual framework for business Biosystems Engineering, 92(2), 143–155.
intelligence based on activities monitoring systems. Sørensen C G; Suomi P; Kaivosoja J; Pesonen L; Norros L (2008).
International Journal of Intelligent Enterprise, 1(1), 65. User-centric information modeling. In: International
Foulds L R; Wilson J M (2005). Scheduling operations for the Conference: September 15–17, 2008 Ragusa, Italy. ‘‘Innovation
harvesting of renewable resources. Journal of Food Technology to Empower Safety, Health and Welfare in
Engineering, 70, 281–392. Agriculture and Agro-food Systems’’.
Fountas S; Wulfsohn D; Blackmore B S; Jacobsen H L; Sørensen C G; Thomsen F T (2006). Synopsis regarding specific
Pedersen S M (2006). A model of decision making and and general requirements on fleet management within arable
information flows for information-intensive agriculture. farming. Available: www.robocluster.dk.
Agricultural Systems, 87(1), 192–210. Van Heijden Rder; Marchau V (2002). Innovating road traffic
Fountas S; Blackmore B S; Vougioukas S; Tang L; Sørensen C G; management by ITS: a future perspective. International
Jørgensen R (2007). Decomposition of agricultural tasks into Journal of Technology, Policy and Management, 2(1), 20–39,
robotic behaviours. Agricultural Engineering International: the 2002.
CIGR Ejournal, Manuscript PM 07 006, IX. Wilson B (2001). Soft Systems Methodology. John Wiley & Sons Ltd.
Gelb E, Offer A (Eds.) (2005). ICT in Agriculture: Perspectives of Witney B (1988). Choosing & Using Farm Mmachines. Longman
Technological innovation. E-book composed under the Scientific & Technical, England. 412 pp.