TOWARDS FOURTH-PARTY LOGISTICS PROVIDERS

A Business Model for Cloud-based Autonomous Logistics

A. Schuldt

, K. A. Hribernik

, J. D. Gehrke

, K.-D. Thoben

and O. Herzog

Centre for Computing and Communication Technologies (TZI), University of Bremen

Am Fallturm 1, D-28359 Bremen, Germany

BIBA - Bremer Institut für Produktion und Logistik GmbH, University of Bremen

Hochschulring 20, D-28359 Bremen, Germany

Keywords:

Agents, Cloud computing, Autonomous control in logistics, Internet of things, Fourth-party logistics.

Abstract:

Cloud computing denotes a paradigm shift in computing that enables a ﬂexible allocation of hardware and

software resources on demand. Therewith, it is particularly appealing for applications with a high degree of

computational complexity and dynamics. This paper identiﬁes logistics planning and control as a promising

application for clouds. However, two prerequisites must be met for cloud-based logistics control. Firstly, the

platform-as-a-service layer must provide a synchronisation of the physically distributed real-world material

ﬂowsand the data ﬂows in the cloud. Secondly, appropriate and scalable control software must be implemented

on the software-as-a-service layer. Apart from outlining the technical foundations, this paper describes how

both steps enable a business model that is usually referred to as fourth-party logistics.

1 INTRODUCTION

The cloud computing paradigm envisions that hard-

ware and software resources are ﬂexibly allocated on

demand. Usually, three different layers are distin-

guished (Rittinghouse and Ransome, 2010):

Infrastructure as a Service refers to a scalable

hardware infrastructure.

Platform as a Service covers a system environment

for application deployment in the cloud.

Software as a Service means that users are provided

with the whole software demanded by them.

Employing clouds raises several questions which

cover, for instance, service reliability, connectivity,

and security. Indisputably, these questions are worth

being addressed. Nevertheless, these questions are

out of the scope of this paper which lays its particular

focus on advantages and resulting use cases.

A promise of cloud computing is that users can

signiﬁcantly decrease their investments for own IT

infrastructure. In principle, only thin clients are re-

quired to access cloud resources through the Internet.

The required computational power is billed based on

the resources actually used. Particularly the ability

to allocate services dynamically distinguishes cloud

computing from application service providing in gen-

eral. Conventional IT infrastructures must be capable

of handling the maximally expected load. Cloud com-

puting approaches this challenge by bringing together

users with complementing demands, i. e., load peaks

of some users coincide with idle times of others (Rit-

tinghouse and Ransome, 2010).

The ability to ﬂexibly adapt to both increasing and

decreasing demands makes cloud computing partic-

ularly appealing for applications that exhibit a high

degree of computational complexity and dynamics.

This particularly holds for logistics planning and con-

trol. As a foundation, characteristics of the logistics

domain are examined and the resulting requirements

are matched to properties of cloud computing (Sec-

tion 2). The ﬁrst requirement is that the software for

logistics control is scalable. To this end, an agent-

based approach for autonomous control in logistics is

presented (Section 3). The second requirement is that

the physically distributed logistics processes are con-

nected to the control system. The necessary synchro-

nisation of material and data ﬂow is accomplished

with Internet of Things technology (Section 4). Com-

bining both aspects with cloud computing allows im-

plementing a business model which is often referred

to as fourth-party logistics, 4PL in short (Section 5).

445

Schuldt A., A. Hribernik K., D. Gehrke J., Thoben K. and Herzog O..

TOWARDS FOURTH-PARTY LOGISTICS PROVIDERS - A Business Model for Cloud-based Autonomous Logistics.

DOI: 10.5220/0003392104450451

In Proceedings of the 1st International Conference on Cloud Computing and Services Science (CLOSER-2011), pages 445-451

ISBN: 978-989-8425-52-2

 2011 SCITEPRESS (Science and Technology Publications, Lda.)

The contribution of this paper is thus a new busi-

ness model for logistics control based on cloud com-

puting. The foundation of this business model is the

integration of agent-based autonomous control in lo-

gistics and cloud computing by means of Internet of

Things technology.

2 CHALLENGES OF LOGISTICS

PLANNING AND CONTROL

Logistics is of considerable importance as a backbone

of the globalised economy. Goods are procured from

all over the world, processed, and distributed to cus-

tomers which can again be located all over the world.

Managing such supply networks is a challenging task

due to three properties:

Complexity which is due to the high number of lo-

gistics objects and their manifold parameters.

Dynamics induced, e.g., by transient customer de-

mands and changes in the environment.

Distribution as logistics processes frequently span

companies, countries, and even continents.

Computing optimal plans that incorporate all relevant

aspects takes a considerable amount of time (Apple-

gate et al., 2007). Due to the dynamics, optimal plans

are therefore often already outdated in the moment

their generation is ﬁnished (Windt and Hülsmann,

2007). This is particularly challenging for logistics

control which requires quick responses to exceptions

that occur while plans are executed.

Given that logistics planning and control can be

executed in parallel, cloud computing is a promis-

ing approach to tackle the challenges caused by com-

plexity and dynamics with increased computational

power. The problem, however, is that optimisation

problems in logistics are usually solved with dis-

crete linear programming which is known to be NP-

complete (Hopcroft and Ullman, 1979). Although

the artiﬁcial Travelling Salesman Problem, as an ex-

ample, might appear simple at ﬁrst glance, it never-

theless exhibits a factorial computational complexity.

Even with sophisticated heuristics, the seemingly in-

exhaustible hardware power of clouds can often not

sufﬁce for solving such problems in acceptable time

(Applegate et al., 2007).

This means that an adequate Software-as-a-

Service layer is required that is capable of reducing

the computational effort and coping with the local

dynamics occurring. Given this prerequisite (which

will be addressed in Section 3), it is possible to bene-

ﬁt from the scalable Infrastructure-as-a-Service layer

of clouds. This is particularly beneﬁcial in order to

tackle global dynamics occurring. These dynamics

may be due to economical or seasonal inﬂuences as

well as special sales concepts. Examples for the for-

mer ones are the ﬁnancial crisis of 2007–2010 as well

as Christmas sales. An example for the latter one are

sales concepts with a continuously changing range of

products (Schuldt, 2010). With cloud computing, lo-

gistics companies would be able to scale their compu-

tational power for process control in accordance with

these effects.

Unlike many other cloud applications, however,

logistics control exhibits another challenge. Logis-

tics control demands frequent interaction with the real

world. On the one hand, this means that the soft-

ware system must be informed about events occur-

ring in the real world. On the other hand, the sys-

tem must be able to react on these changes, i. e., its

decisions must be executed in the real world. The

required synchronisation of material and data ﬂows

is particularly challenging because logistics processes

are inherently distributed. This high degree of spa-

tial distribution prevents local information from be-

ing available for central decision-making (Jedermann

and Lang, 2008). This is part of the Platform-as-

a-Service layer (which is dealt with in Section 4),

which connects the Infrastructure-as-a-Service and

the Software-as-a-Service layers.

To summarise the ﬁndings so far, the requirements

for cloud-based logistics control are as follows:

1. On the Software-as-a-Service layer, it is necessary

to implement an adequate control software that

is capable of coping with both the computational

complexity and dynamics.

2. On the Platform-as-a-Service layer, it is necessary

to synchronise real-world material ﬂows with the

data ﬂows in the cloud to account for the physical

distribution of logistics processes.

3 AGENT-BASED AUTONOMOUS

CONTROL IN LOGISTICS

As elaborated in the previous section, conventional

approaches to logistics planning are not applicable

for on-line control. Due to their asymptotic compu-

tational complexity, even the computational power of

clouds does not sufﬁce (Section 2). A particular chal-

lenge is the dynamics of logistics processes that ren-

ders optimal plans outdated in the moment their gen-

eration is ﬁnished. Think, for instance, of a shipping

container loaded with perishable goods. The shelf

life of the goods decreases if the interior tempera-

ture increases. If a sensor network within the con-

CLOSER 2011 - International Conference on Cloud Computing and Services Science

446

tainer detects such an increase in temperature, it is

thus necessary to re-route the container to another lo-

cation nearby and to send another container with sim-

ilar goods to the original destination. Other reasons

for re-planning include trafﬁc or weather conditions.

All these reasons for dynamics have in common that

they only affect some logistics objects and by far not

the whole network.

Motivated by this ﬁnding, the paradigm of au-

tonomous control in logistics delegates decision-

making to the participating logistics objects them-

selves (Windt and Hülsmann, 2007). Autonomous

control enables logistics objects to process informa-

tion, to make and execute decisions, and to cooperate

with each other based on objectives imposed by their

owners. The advantages over centralistic approaches

are as follows (Schuldt, 2010):

1. The computational effort is signiﬁcantly reduced

by computational decomposition.

2. The scalability of process control is signiﬁcantly

increased by parallelising decision-making.

3. Reactivity and robustness are signiﬁcantly in-

creased by local exception handling.

Potential autonomous logistics entities are compo-

nents, articles, sales units, cardboard boxes, pallets,

and shipping containers.

The technologies enabling autonomous logistics

are identiﬁcation, localisation, sensors, communi-

cation, as well as local data processing. Fig-

ure 1 depicts an architecture for autonomous lo-

gistics entities (Schuldt, 2010). The identiﬁcation

unit uniquely identiﬁes autonomous logistics entities

(Hribernik et al., 2009). The localisation unit al-

lows self-localisation of logistics entities, e. g., based

on global navigation satellite systems (Hofmann-

Wellenhof et al., 2008). The sensor unit continuously

monitors the environmental and the interior state of

the logistics object (Al-Karaki and Kamal, 2004; Jed-

ermann and Lang, 2008). The communication unit

enables coordination with other entities. The heart

of autonomous logistics entities, however, is the data

processing unit which checks whether the original

planning for the object is still valid or whether it

must be updated. Its decision-making can be imple-

mented by means of intelligent software agents. Ex-

ample applications of agent-based autonomous logis-

tics cover adaptive truck routing incorporatingdriving

time estimation based on trafﬁc and weather condi-

tions (Gehrke and Wojtusiak, 2008) and autonomous

container dispatch (Schuldt, 2010).

A physically distributed application of software

agents is possible (Adorni et al., 2001). For instance,

this makes sense in order to continuously monitor

Identification

Unit

Localisation

Unit

Sensor

Unit

Data Processing Unit

Communication Unit

Autonomous Logistics Entity

Other Participant

Autonomous

Logistics Entity

Environment

Figure 1: Architecture for autonomous logistics entities.

sensor measurements without the necessity to trans-

mit them over communication networks. Embedded

systems attached to the logistics objects, however, are

not designed for extensive reasoning tasks. Instead,

clouds are a promising platform for such re-planning.

Agent-based implementations natively support

parallel execution. In contrast to sequentially exe-

cuted software solutions for logistics control, soft-

ware agents are thus well prepared for virtualisa-

tion. Multiagent platforms such as JADE (Bellifem-

ine et al., 2007) support multiple so-called agent con-

tainers in which agents can be deployed at differ-

ent locations. Agents can then communicate with

each other also over the boundaries of their particu-

lar agent container. Interoperability can be ensured

because the agent concept has been shown to be

sufﬁciently close to the service-oriented architecture

(SOA) model (Moreau, 2002). A number of con-

cepts exist which interface multiagent platforms with

SOA middleware, such as WS2JADE (Nguyen and

Kowalczyk, 2007), middle agents (Sycara, 2001), and

other gateway architectures which allow transparent,

fully automatic interoperation (Greenwood and Cal-

isti, 2004).

In order for cloud-based autonomous logistics to

be operationalised, however, an adequate approach

must be developed to synchronise the material and

data ﬂows in the respective logistics processes. The

actual logistics objects – whether they be physical

such as goods, containers, trucks and transport hubs

or immaterial such as orders – need to be connected

both to the agents which are their digital counterparts

in the cloud and to existing logistics IT infrastruc-

ture such as dispatch, route planning, and enterprise

resource planning systems in order for cloud-based

logistics service provision to be feasible. In terms

TOWARDS FOURTH-PARTY LOGISTICS PROVIDERS - A Business Model for Cloud-based Autonomous Logistics

447

of the architecture for autonomous logistics entities

(Figure 1), this means that the data processing unit

which is then located in the cloud must be connected

to all other units.

4 MATERIAL AND DATA FLOW

SYNCHRONISATION

To summarise the ﬁndings thus far, agent-based au-

tonomous logistics helps cope with the computa-

tional complexity and dynamics on the Software-as-a-

Service layer (Section 3). Another important prereq-

uisite (Section 2), however, is the synchronisation of

material and data ﬂows on the Platform-as-a-Service

layer. In particular, three challenges have to be ad-

dressed:

Unique identiﬁcation of logistics objects to estab-

lish a link between real world objects and their

agent counterparts.

Data integration from varioussources in a meaning-

ful manner.

Dynamic data source integration such as RFID,

sensors, sensor networks, and other systems in-

tegrated in logistics objects.

For these purposes, the concept of the Internet of

Things becomes relevant. In essence, the Internet of

Things extrapolates the idea of the Internet – a global,

interconnected network of computers – to describe

a network of interconnected things, such as every-

day objects, products, and environments. As such,

the concept represents the convergence of a num-

ber of recent multi-disciplinary developments such as

Ambient Intelligence (Ducatel et al., 2001), Ubiqui-

tous (Weiser, 1991) and Pervasive Computing (Gupta

et al., 2001), Auto Identiﬁcation (Cole and Engels,

2002), and Intelligent Products (Meyer et al., 2009).

At the heart of the concept lies the idea that objects

– things – are capable of information processing and

communication with each other and with their envi-

ronment. For cloud-based logistics control, the logis-

tics objects are the things which the multiagent sys-

tem enables to process information and communicate.

An important prerequisite for autonomous logis-

tics based on cloud computing is its integration into

existing logistics infrastructures. Therefore, it is im-

portant to synchronise real-world material ﬂows and

data ﬂows in the cloud. This mapping can be accom-

plished based on the identiﬁcation standards of the

EPCglobal Architecture Framework (Hribernik et al.,

2009). ID@URI using the Dialog system (Främling

et al., 2006) is an alternative that combines unique

article identiﬁers with Internet addresses where addi-

tional information about the object can be retrieved.

Furthermore, it is necessary to integrate data from

various sources in a semantically meaningful manner.

To this end, semantic mediators can be applied. Fig-

ure 2 illustrates a concept for a generic data integra-

tion concept for logistics clouds with the Internet of

Things. It is an extension of a concept for generic

data integration in autonomous logistics (Hribernik

et al., 2010) which satisﬁes requirements both to-

wards the coupling of the material and data ﬂows as

well as towards providing uniﬁed access to all rele-

vant data sources including interfacing to the service

consumers. At the heart of the concept lies a medi-

ator component (Ullman, 1997; Wache et al., 2001),

which is capable of composing queries to any combi-

nation of relevant logistics data sources. It achieves

this by semantic mediation. Each data source is fully

described syntactically and semantically by an ontol-

ogy, which can be mapped onto the others by the me-

diator. Wrapper components handle the transforma-

tion to and from the relevant data sources in a rule-

based fashion. By implementing semantic descrip-

tions and transformation rules for widely used lo-

gistics data exchange formats such as the EDIFACT

subset EANCOM or standards from the EPCglobal

Architecture Framework such as EPCIS (Electronic

Product Code Information Services) (EPCglobal Inc.,

2007), access to the majority of relevant data sources

is given. Additional, proprietary data sources can

be integrated simply by adding a new wrapper with

the relevant semantic description and set of transfor-

mation rules, making the concept highly extensible.

This approach also allows the service consumer to ei-

ther easily integrate the required services into its own

logistics IT landscape, or, for example, utilise thin

clients to access a web-based GUI towards the cloud

services.

Finally, the proposed concept also facilitates the

direct integration of dynamic data sources used in lo-

gistics processes, such as such as RFID, sensors, sen-

sor networks and other systems integrated into phys-

ical logistics objects. By abstracting from the physi-

cal interfaces towards these data sources, the seman-

tic mediation approach may be applied in much the

same way it is to static data sources. The abstraction

layer is required to be able to provide a reliable in-

terface regardless of the physical accessibility of the

dynamic data sources at any time. It is responsible

for buffering, ﬁltering and routing data to and from

the respective data sources. It may consist of ele-

ments such as the FOSSTRAK HAL towards EPC-

compliant RFID (Floerkemeier et al., 2007), PMI

(Promise Messaging Interface) (Främling and Ny-

CLOSER 2011 - International Conference on Cloud Computing and Services Science

448

Enterprise Systems

(ERP, WMS, TMS, etc.)

EPCglobal Architecture Framework

(EPCIS, ONS, etc.)

EDI and Messaging

(EDIFACT EANCOM etc.)

Proprietary

Interfaces

Multiagent System of

Autonomous Logistics Entities

Data Processing Unit

RFID Sensors

Hardware Abstraction Layer

Semantic Descriptions and Transformation Rules

Ontology-Based Mediator

Generic Integration Services

Embedded Systems

Semantic Data Integration

Logistics Data Sources

Standard Logistics Interfaces

Data Processing UnitData Processing Unit

Autonomous Logistics Entity Autonomous Logistics Entity Autonomous Logistics Entity

Figure 2: Data integration concept for cloud-based autonomous logistics entities.

man, 2008) towards PEIDs (Product Embedded In-

formation Devices) (Kiritsis et al., 2003) or OSGi to-

wards sensor components (Ahn et al., 2006).

5 BUSINESS MODEL FOR

FOURTH-PARTY LOGISTICS

Despite all of the advantages cloud computing offers,

the broad variety of involved technologies makes it

challenging for individual logistics companies to ben-

eﬁt from these advantages. Logistics enterprises are

experts in logistics and not in IT. Employing spe-

cialised service providers that offer logistics cloud

computing is thus preferable. This leads to different

business models for service providers.

5.1 Logistics Cloud Service Providers

The ﬁrst three layers of logistics cloud computing

correspond more or less directly to the three general

cloud layers Infrastructure, Platform, and Software as

a Service (Section 1). Nevertheless, service providers

for logistics clouds have to ensure the services out-

lined in Sections 3 and 4. However, it is worth men-

tioning that service providers on different layers may

be disjoint from each other. That is, providers on

higher levels may acquire services on lower levels

from other service providers.

On the Infrastructure-as-a-Service layer, there is a

direct correspondence between logistics cloud com-

puting and general cloud computing as both refer

to the underlying hardware infrastructure. Conse-

quently, service providers on this level do not need

any expertise in logistics. On the contrary, it seems

even advantageous if logistics companies share com-

puting resources with enterprises from completely

different branches, at least given that their computa-

tional demands complement each other. This helps

utilising the resources of the infrastructure service

provider more efﬁciently.

Logistics-speciﬁc tasks on the Platform-as-a-

Service layer can be categorised as follows. Firstly,

fundamental services for agent deployment in the

cloud have to be provided (Section 3). This eases the

deployment of agent representatives and delegates the

administration of the software platform to the service

provider. Secondly, the synchronisation of material

and data ﬂows has to be implemented in order to es-

tablish the link between real-world logistics objects

and their digital counterparts in the cloud (Section 4).

For data integration, they provide interfaces to data

sources and mediate if necessary. Logistics compa-

nies do thus not have to consider these issues them-

selves and can thus implement their logistics control

without being burdened with data integration.

The Software-as-a-Service layer of logistics

clouds goes even one step further. Service providers

offer a complete implementation of the software

agents needed for autonomous control of a speciﬁc

process. Furthermore, the administration of software

agents is left to the service provider. Consequently,

service providers on this layers need expertise in the

ﬁeld of logistics control with multiagent systems. In

return, the need for IT expertise in the logistics com-

pany is reduced to a minimum because its task re-

duces to delivering relevant process information.

5.2 Fourth-party Logistics Providers

On the Software-as-a-Service layer, the cloud

provider offers the software to control logistics pro-

TOWARDS FOURTH-PARTY LOGISTICS PROVIDERS - A Business Model for Cloud-based Autonomous Logistics

449

cesses. The Platform-as-a-Service layer provides the

means to interconnect the IT systems with the real-

world logistics objects. Logistics enterprises are

thus relieved from concerning the IT infrastructure.

Instead, their task reduces to ensuring that the re-

quired logistics resources and information are avail-

able. Apart from the cloud infrastructure, enterprises

thus have to make framework contracts with logistics

service providers, e. g., about transport and storage re-

sources. These are then available to the agent repre-

sentatives of logistics objects in the cloud. For in-

stance, an agent can initiate a transport for a shipping

container it represents based on the pre-negotiated

framework contract.

The service of the ﬁrst three layers can even be

improved by also providing direct access to logis-

tics resources, i. e., the pre-negotiated framework con-

tracts for logistics resources become superﬂuous. To

this end, the cloud service provider additionally offers

an electronic marketplace on which logistics service

providers and consumers can negotiate (Smith, 1977)

on logistics services:

1. An agent wants to allocate logistics resources for

its respective logistics object.

2. Matching logistics service providers are discov-

ered by means of semantic service descriptions.

3. A negotiation between the service consumer agent

and the agents of the potential service providers

takes place.

4. The best offer is chosen.

This means that also the procurement of logistics ser-

vices is moved into the cloud. Logistics customers

thus get a ready-to-use solution for integrated logis-

tics control. In this business model, it is not neces-

sary to bill customers for the utilisation of cloud re-

sources explicitly. Each transaction in the cloud is di-

rectly linked with a transaction of logistics resources.

Billing for cloud services can thus be directly cou-

pled with billing for the logistics services. The costs

for computing will usually be negligible compared to

the negotiated logistics services like transport.

A cloud service provider acting on this level might

become what is often referred to as a fourth-party lo-

gistics provider, 4PL in short. Fourth-party logistics

providers are considered to have no own logistics re-

sources. Instead, they employ sophisticated IT sys-

tems in order to integrate services from freight opera-

tors and stockists, forwarding agencies, and outsourc-

ing companies in the ﬁeld of contract logistics.

6 CONCLUSIONS AND

OUTLOOK

Logistics control is a promising ﬁeld of application

for cloud computing. Logistics processes exhibit both

a high degree of complexityand dynamics. Therefore,

a scalable IT infrastructure is required for supply net-

work management. The computational complexity of

conventional approaches to logistics planning makes

them inapplicable for on-line control. By contrast,

the paradigm in agent-based autonomous control in

logistics can serve as the Software-as-a-Service layer

of logistics clouds. In this approach, decision-making

is delegated to agent representatives of the individual

logistics objects. For this purpose, the synchronisa-

tion of material ﬂows in highly distributed logistics

processes and their corresponding data ﬂows in the

cloud is another challenge. It can be accomplished

with Internet-of-Things technology which is imple-

mented on the Platform-as-a-Service layer.

Providing each of the above services is a business

model on its own. Both relieve logistics enterprises

from IT-related tasks, thus enabling them to focus on

their core business. From their combination, however,

emerges another promising business model: fourth-

party logistics providers who do not have own logis-

tics resources but combine services of other providers

to custom-tailored logistics.

The investigation so far shows that cloud-based

autonomous control actually implies promising busi-

ness models (Section 5). The feasibility of imple-

menting the Platform-as-a-Service (Section 4) and

the Software-as-a-Service (Section 3) layer has been

shown in this paper. The next step following this fea-

sibility study is thus to actually implement cloud com-

puting for logistics control.

ACKNOWLEDGEMENTS

This research is funded by the German Research

Foundation (DFG) within the Collaborative Research

Centre 637 “Autonomous Cooperating Logistic Pro-

cesses: A Paradigm Shift and its Limitations”

(SFB 637) at the University of Bremen, Germany.

REFERENCES

Adorni, G., Bergenti, F., Poggi, A., and Rimassa, G. (2001).

Enabling FIPA Agents on Small Devices. In CIA

2001, pages 248–257, Modena, Italy. Springer-Verlag.

Ahn, H., Oh, H., and Sung, C. O. (2006). Towards Reliable

CLOSER 2011 - International Conference on Cloud Computing and Services Science

450

OSGi Framework and Applications. In SAC 2006,

pages 1456–1461, Dijon, France. ACM Press.

Al-Karaki, J. N. and Kamal, A. E. (2004). Routing Tech-

niques in Wireless Sensor Networks: A Survey. IEEE

Wireless Communications, 11(6):6–28.

Applegate, D. L., Bixby, R. E., Chvátal, V., and Cook, W. J.

(2007). The Traveling Salesman Problem: A Compu-

tational Study. Princeton University Press.

Bellifemine, F., Caire, G., and Greenwood, D. (2007). De-

veloping Multi-Agent Systems with JADE. John Wiley

& Sons, Chichester, UK.

Cole, P. H. and Engels, D. W. (2002). Auto ID – 21st Cen-

tury Supply Chain Technology. In AEEMA 2002, Mel-

bourne, Australia. AEEMA.

Ducatel, K., Bogdanowicz, M., Scapolo, F., Leijten, J., and

Burgelman, J.-C. (2001). Scenarios for Ambient In-

telligence in 2010. Technical report, European Com-

mission.

EPCglobal Inc. (2007). EPC Information Services (EPCIS)

Version 1.0.1 Speciﬁcation. Standard Speciﬁcation,

EPCglobal Inc.

Floerkemeier, C., Lampe, M., and Roduner, C. (2007). Fa-

cilitating RFID Development with the Accada Proto-

typing Platform. In PerCom 2007, pages 495–500,

White Plains, NY, USA. IEEE Computer Society.

Främling, K., Ala-Risku, T., Kärkkäinen, M., and Holm-

ström, J. (2006). Agent-Based Model for Managing

Composite Product Information. Computers in Indus-

try, 57(1):72–81.

Främling, K. and Nyman, J. (2008). Information Architec-

ture for Intelligent Products in the Internet of Things.

In NOFOMA 2008, pages 221–232, Helsinki, Finland.

Gehrke, J. D. and Wojtusiak, J. (2008). Trafﬁc Prediction

for Agent Route Planning. In ICCS 2008), pages 692–

701, Kraków, Poland. Springer-Verlag.

Greenwood, D. and Calisti, M. (2004). Engineering Web

Service – Agent Integration. In SMC 2004, pages

1918–1925, The Hague, The Netherlands. IEEE.

Gupta, S. K. S., Lee, W.-C., Purakayastha, A., and Srimani,

P. K. (2001). An Overview of Pervasive Computing.

IEEE Personal Communications, 8(4):8–9.

Hofmann-Wellenhof, B., Lichtenegger, H., and Wasle, E.

(2008). GNSS Global Navigation Satellite Systems:

GPS, GLONASS, Galileo and more. Springer-Verlag,

Wien, Austria.

Hopcroft, J. E. and Ullman, J. D. (1979). Introduction

to Automata Theory, Languages, and Computation.

Addison-Wesley, Reading, MA, USA.

Hribernik, K. A., Hans, C., and Thoben, K.-D. (2009). The

Application of the EPCglobal Framework Architec-

ture to Autonomous Control in Logistics. In LDIC

2009, Bremen, Germany. Springer-Verlag.

Hribernik, K. A., Kramer, C., Hans, C., and Thoben, K.-

D. (2010). A Semantic Mediator for Data Integration

in Autonomous Logistics Processes. In IESA 2010,

London, UK. Springer-Verlag.

Jedermann, R. and Lang, W. (2008). The Beneﬁts of

Embedded Intelligence – Tasks and Applications for

Ubiquitous Computing in Logistics. In IOT 2008,

pages 105–122, Zurich, Switzerland. Springer-Verlag.

Kiritsis, D., Bufardi, A., and Xirouchakis, P. (2003). Re-

search Issues on Product Lifecycle Management and

Information Tracking Using Smart Embedded Sys-

tems. Advanced Engineering Informatics, 17(3–

4):189–202.

Meyer, G. G., Främling, K., and Holmström, J. (2009). In-

telligent Products: A Survey. Computers in Industry,

60(3):137–148.

Moreau, L. (2002). Agents for the Grid: A Comparison

with Web Services (Part I: Transport Layer). In CC-

GRID 2002, pages 220–228, Berlin, Germany. IEEE

Computer Society.

Nguyen, X. T. and Kowalczyk, R. (2007). WS2JADE: In-

tegrating Web Service with Jade Agents. In SOCASE

2007, pages 147–159, Honolulu, HI, USA. Springer-

Verlag.

Rittinghouse, J. W. and Ransome, J. F. (2010). Cloud Com-

puting: Implementation, Management, and Security.

CRC Press, Taylor & Francis, Boca Raton, FL, USA.

Schuldt, A. (2010). Multiagent Coordination Enabling Au-

tonomous Logistics. Doctoral dissertation, Universität

Bremen.

Smith, R. G. (1977). The Contract Net: A Formalism for

the Control of Distributed Problem Solving. In IJ-

CAI 1977, page 472, Cambridge, MA, USA. William

Kaufmann.

Sycara, K. P. (2001). Multi-Agent Infrastructure, Agent

Discovery, Middle Agents for Web Services and In-

teroperation. In EASSS 2001, pages 17–49, Prague,

Czech Republic. Springer-Verlag.

Ullman, J. (1997). Information Integration Using Logical

Views. In ICDT 1997, pages 19–40, Delphi, Greece.

Springer-Verlag.

Wache, H., Vögele, T., Visser, U., Stuckenschmidt, H.,

Schuster, G., Neumann, H., and Hübner, S. (2001).

Ontology-Based Integration of Information - A Sur-

vey of Existing Approaches. In IJCAI 2001 WS on

Ontologies and Information Sharing, pages 108–117,

Seattle, WA, USA.

Weiser, M. (1991). The Computer for the Twenty-First Cen-

tury. In Scientiﬁc American, volume 265, pages 94–

104.

Windt, K. and Hülsmann, M. (2007). Changing Paradigms

in Logistics. In Hülsmann, M. and Windt, K., editors,

Understanding Autonomous Cooperation and Control

in Logistics, pages 1–16. Springer-Verlag, Heidelberg,

Germany.

TOWARDS FOURTH-PARTY LOGISTICS PROVIDERS - A Business Model for Cloud-based Autonomous Logistics

451