379
Chapter 14
Toward Semantics-Based
Service Composition
in Transport Logistics
Joerg Leukel and Stefan Kirn
Universität Hohenheim, Stuttgart, Germany
Contents
14.1 Introduction .............................................................................................380
14.2 Basic Assumptions ....................................................................................381
14.2.1 Logistics System ............................................................................381
14.2.2 Transport Service ..........................................................................381
14.2.3 Composite Transport Service ........................................................382
14.3 Semantic Model for Service Composition ................................................383
14.3.1 Semantics of Logistics System .......................................................383
14.3.2 Semantics of Nodes (N) ................................................................383
14.3.2.1 Semantics of Flows (F) ....................................................384
14.3.2.2 Semantics of Transportation Means................................384
14.3.2.3 Semantics of Transport Units (TUs) ...............................384
14.3.2.4 Semantic Integrity of TM and TU .................................385
14.3.3 Semantics of Transport Service .....................................................386
14.3.3.1 Semantics of Service Flows (S) ........................................386
14.3.3.2 Semantics of Actors ........................................................386
14.3.3.3 Semantic Integrity of M ..................................................386
14.3.3.4 Semantics of Composite Transport Services (CS) ...........387
380 ◾ Joerg Leukel and Stefan Kirn
14.1 Introduction
Logistics is a domain concerned with controlling and executing the ow of goods,
services, and associated information from sources to destinations, e.g., from manu-
facturing site to point of sale. It can be characterized by multiple rms providing
resources and services or delivering complex services to meet customer require-
ments. e need for coordination across rms is obvious, since few single rms
deliver an entire product or service without contractors because they face division
of labor for basic logistics services, increasing customer requirements that lead to
greater specialization, and supply chain management (SCM).
e task of a logistics system is to transform goods with regard to location,
time, and quantity. ese transformations materialize into the concept of logistics
service—a logical set of transformations. Logistics services are oered by rms such
as shippers, packers, warehouses, and rms that provide more complex services. In
this sense, a logistics system provides capabilities to deliver services to customers
but does not dene a priori the ways to implement the services.
e problem with logistics systems is nding the best solution for a given set of
customer requirements. We will dene this problem as a subclass of service composi-
tion, thus combining and linking services. e result is a composite logistics service.
Composition has two dimensions: hierarchy (part-of relationships between services) and
sequence (logical order of services). In prior research, we studied the mapping of logis-
tics services to models, methods, and technologies of Web Service research (Karaenke
and Kirn 2009) in particular, by employing languages for describing service level agree-
ments (SLAs) formally dening obligations and guarantees in a service relationship.
We address the composition problem from the perspective of interoperability;
hence, we aim to make the semantics of all relevant parts of logistics systems
explicit, machine-readable, and exchangeable. Such a semantic description is based
on the annotation principle, as adopted in Semantic Web services (Martin et al.,
2007) by which service providers maintain their local descriptions while annotat-
ing them according to a shared conceptualization (ontology). We think that this
approach is feasible in specic environments, such as regional logistics networks of
shippers. We limit the scope to transport logistics and thus transport services.
Current representations of such services are not semantic. ey can be found
in intra-organizational information systems such as ERP and inter-organizational
14.4 Preliminary Validation .............................................................................388
14.4.1 Validation Scenario .......................................................................388
14.4.2 Semantic Modeling .......................................................................388
14.5 Discussion ................................................................................................390
14.6 Related Work ...........................................................................................391
14.7 Conclusions ..............................................................................................392
Acknowledgement .............................................................................................392
References .........................................................................................................393
Toward Semantics-Based Service Composition ◾ 381
information systems such as logistics marketplaces and SCM. We enrich these rep-
resentations by using description logic (DL) that allows for reasoning about cus-
tomer requirements and answering more expressive queries during service discovery
and composition. ese are key features, because customer requirements are often
expressed in an abstract form—searching for transportation of hazardous goods
without specifying vehicle types and load-securing measures.
e objectives of this chapter are to (1) develop a semantic model for transport
services and (2) apply the model to a use case scenario to demonstrate its feasibil-
ity and usefulness. e next section denes the basic assumptions relevant to the
domain and problem. After that, we present our semantic model, followed by a
section on preliminary validation and discussion. We also review related work and
nally draw conclusions and outline avenues of future research.
14.2 Basic Assumptions
is section denes the basic terms and formalisms for logistics systems, transport
services, and composite transport services.
14.2.1 Logistics System
Denition 1: A logistics system consists of nodes participating in transforming
goods with regard to location, time, and quantity. e inter-relations between
nodes are constituted by the possible ow of goods (e.g., modes of transportation).
Nodes represent the storage of goods at locations such as warehouses. A logistics
system model is represented by a directed graph LS = (N, F), where N is the set of
all nodes and F is the set of all possible ows of goods with F ⊆ N × N × TM × TU.
Each f is a four-tuple, f = (n
j
, n
k
, tm, tu), with ow from n
j
to n
k
using a transporta-
tion means tm ∈ TM (e.g., truck) and transport unit tu ∈ TU (e.g., container). e
following integrity constraints must hold:
Let •n = {m|(m,n)∈F}, hence the set of input nodes of n; then at least one n ∈ N
exists with |•n| = 0. us, at least one node has no incoming ows, i.e., representing
origin of goods.
Let n• = {m|(n,m)∈F}, hence the set of output nodes of n; then at least one n ∈ N
exists with |n•| = 0. us, a least one node has no outgoing ows, i.e., representing
nal destination of goods.
For all n ∈ N: |n•|+|•n|≥1. us, the graph LS is (weakly) connected.
14.2.2 Transport Service
Denition 2: A transport service is a possible service ow from a logistics service
provider to a logistics service consumer. e set of all such services forms a trans-
port service ow model, which is a directed graph SF = (A, S, M). A is the set of
all actors. S is the set of all possible elementary service ows. Each s ∈ S is a tuple
382 ◾ Joerg Leukel and Stefan Kirn
s = (a
j
, a
k
), with service ow from a
j
to a
k
. M is a relation that maps each s to one
or more f in LS. us each service ow s relates to |M(s)| ows of goods in L. e
following integrity constraints must hold:
Let •a = {b|(b,a)∈S}, then at least one a ∈ A exists with |•a|= 0, i.e., actor that is
not a logistics service requester.
Let a• = {b|(a,b)∈S}, then at least one a ∈ A exists with |a•|=0, i.e., actor that is
not a logistics service provider.
SF is (weakly) connected.
For any s, if |M(s)|≥2, then t = |M(s)|, and there must exist a walk w
s
in LS with
w
s
(n
i
, m
1
, …, m
t
, n
j
) and n
i
, n
j
∈ N.
14.2.3 Composite Transport Service
Denition 3: A composite transport service is a composition of two or more
transport services in SF. It is dened as a four-tuple cs = (a
j
, a
k
, CE, WF). e
composite service is oered by actor a
j
∈ A to actor a
k
∈ A. CE denotes the com-
position elements. WF denotes the workow, i.e., logical sequence of all ce ∈ CE.
e following integrity constraints must hold:
Let CS be the set of all composite services. A composite service cs may contain
elementary and other composite services, i.e., ∀ cs
i
: CE ⊆ S* ∪ CScs
i
. WF describes
a directed tree (A’, S’) with A’ ⊆ A, S’ ⊆ S.
Figure14.1 illustrates the inter-relationships of these formalisms: LS represents
a two-stage logistics system with transportation of goods from three sources via a
Composite service
a
3
a
4
cs
Logistics service
flow model
s
1
Abstraction
by Services
Abstraction
by Composition
a
3
a
1
SF = (A, S, M)
a
2
s
2
n
1
n
2
n
3
n
5
n
4
Logistics system
model
LS = (N, F)
f
1
f
2
f
3
f
4
cs = (a
3
, a
4
, CE, WF)
CE = {s
1
, s
2
}
Figure 14.1 Examples of logistics system, service flow, and composite service.
Toward Semantics-Based Service Composition ◾ 383
hub (n
4
) to the nal destination (n
5
). ere are two service providers: a
1
oering
transportation from sources to the hub and a
2
oering transportation to the desti-
nation site; these services are consumed by a
3
. In addition, the actor a
4
is a so-called
third party logistics provider (3PL) oering a composite service combining both
services, and therefore covers all transportation relationships in LS.
14.3 Semantic Model for Service Composition
is section develops the semantic model for transport services. It augments the
logistics system and service ow model with additional formal semantics of each
element and discusses classications and constraints of each element. A semantic
description may be grounded on respective domain ontologies, but viable ontolo-
gies for logistics do not provide explicit, formal, and machine-readable representa-
tion (Leukel and Kirn 2008). erefore, we provide a core conceptualization of the
domain based on logistics systems and SCM theory (Storey et al. 2006). We dene
the semantics using the SHOIN description logic due to its expressiveness (Baader
et al. 2007); SHOIN is also the underlying logic of the OWL DL Web ontology
language (W3C 2004).
14.3.1 Semantics of Logistics System
14.3.2 Semantics of Nodes (N)
e informal meaning is that each node represents a physical location where
goods can be stored for a limited time. e nodes can be classied by several
methods.
Geographical–political classication—One can adopt ISO 3166 dening
standard codes for all country names (ISO 2010). However, this classication is
at, with disjunction of all classes as the only semantics. us, we dene a func-
tional role (N, Country):locatedIn with T ⊑ 1locatedIn and add assertions for all ISO
countries by respective individual, e.g., {CAN}:Country for Canada, {FRA}:Country
for France, etc., with {CAN} ⊑ {FRA}, etc. e decision to choose individuals
over subconcepts is made because each country exists only once (e.g., there is no
instance of a Sweden concept). We could also incorporate additional information
about countries into the ontology by adding roles and making assertions about
those individuals.
Geographical–logistics classication—e United Nations Code for Trade
and Transport Locations (UN/LOCODE) is a classication specically designed
for the domain (UNECE 2010). It (1) adds a second level to ISO 3166 consist-
ing of about 42,000 locations, (2) denes the logistics functions of each location
(port, rail terminal, road terminal, airport, postal exchange oce, intermodal
transfer, border crossing), and (3) gives the respective geographical coordinates.
..................Content has been hidden....................
You can't read the all page of ebook, please click here login for view all page.