ENTERPRISE SYSTEMS CONFIGURATION AS AN
INFORMATION LOGISTICS PROCESS
A Study
Mats Apelkrans and Anne Håkansson
Jönköping International Business School, Dept of Informatics, Jönköping, Sweden
Dept of Information Science, Computer Science, Kyrkogårdsgatan 10, Uppsala, Sweden
Keywords: Visualisation, Knowledge modelling, Information Logistics, Enterprise Systems, Configuration, Unified
Modeling Language, Verification and Validation
Abstract: In this paper, we suggest using rule-based descriptions of customer’s requirements for Enterprise Systems
implementing Information Logistics. The rules are developed from the users’ requirements and inserted as
schedules to the Enterprise System. The output, from testing these rules, is a list of modules and parameter
settings to configure the system. By using rules, we can, at least partly, automate the configuration process
by traversing the several modules and thousands parameters that are in an Enterprise System. From the list,
we can select the modules and the parameters that meet the customer’s requirements. Then these selected
modules and parameters are visually presented through a kind of Unified Modeling Language diagrams, to
support the user investigation and then to configure the system either manually or automatically. Every
attempt to match a customer’s requirement to the contents of the knowledge base within the Enterprise
system can be thought of as an Information Logistics Process. The output from such a process must be
examined by the user, which can give rise to a new call to the Information Logistics process. In other words
the configuration work is done through a dialogue between the customer and the knowledge base of the
Enterprise system.
1 INTRODUCTION
One of the largest problems in Business Informatics
is to configure chosen Enterprise System (ES)
software to meet a customer’s specific requirements.
This can be seen as a kind of customization of the
ES software. However, following Summer (2005)
and Hedman and Kalling (2002) we will call it
configuration. According to Hedman and Kalling
(2002) “the actual process of implementing an ERP
system is complex. Every ERP system has to be
configured. Configuration involves adapting the
generic functionality of the package to meet the
needs of a particular organization (usually by
setting parameters in tables), such as what calendar
should be applied for a firm with locations in several
countries. All in all, the configuration might affect
thousands of configuration tables and can take years
to complete.” Methods and discussions of
configuring enterprise systems are also found in
(Keller and Teufel 1998; Olson 2003). Summer
(2005) talks about Configuration Management.
Configuration Management provides product
specific configuration support for companies that
must build products for their customers. Technology
vendors, appliance vendors, and computer vendors
are examples of companies, which need to create
product configurations, and make price quotes. A
software package can contain a number of modules
each with a great number of parameters needing to
be set for optimal values. The problems are how to
find them and how to give them the best possible
values. The configuration cannot 100% meet the
requirements, because some modules can be
cancelled and there can be functionalities missing in
the package. In that case, one may have to add extra
software. This will give rise to another problem:
integration of this added software to the original ES
package.
The configuration efforts are both time-
consuming and costly (Adam and Sammon, 2004).
There is also a lack of human competence. People
working with ES configuration usually have only a
212
Apelkrans M. and Håkansson A. (2007).
ENTERPRISE SYSTEMS CONFIGURATION AS AN INFORMATION LOGISTICS PROCESS - A Study.
In Proceedings of the Ninth International Conference on Enterprise Information Systems - ISAS, pages 212-220
DOI: 10.5220/0002370602120220
Copyright
c
SciTePress
partial knowledge of an ES packages’ many
parameters.
In this paper, we concentrate on how to find the
modules and the parameters from the customers’
requirements on the enterprise system. We present a
study on how to partly automate the configuration
process using rule-based descriptions of customer’s
requirements and the modules of the ES package.
These rule-based descriptions are stored in a
Knowledge Base.
The use of production rules on top of the content
of the ES is necessary to keep track of the order of
the tasks performed by the system. The rules handle
all the modules and the parameters used in systems.
With rules, for example, it can be assured that the
calculation of total price in not performed before
having the information about both the price per
product and the quantity of that product. The rules in
the system’s knowledge base already have the order
needed for different tasks but the customers’
requirements may not meet these. Therefore, we
match the rules of the knowledge base with the rules
developed from the customers’ requirements. The
matching between requirements and ES package will
be performed as an Information Logistic Process
(ILP). Hence, there will be a number of ILP calls.
This is the outer loop. The user, through a dialogue,
controls the operation of this loop.
In section 2 and 3 we define the information
logistics process including three phases of the
process. In section 4, we show the common modules
and parameters and in section 5 we present rules
from users’ requirments. The last section includes
matching the users’ requirements to the system’s
rule description, the result of applying rules and
verification of the knowledge base.
2 INFORMATION LOGISTICS
PROCESS
In this paper, we discuss Information Logistics from
an information logistics process perspective: An
information logistics process transforms a given
input into some form of output. The input is some
kind of fragmented information or knowledge
description, which derives from a so-called
information supplier. This input information can be
handled either manually or automatically by the
system. The process output is an information
product that will be made accessible and delivered to
the information receiver who will use the
information. This workflow is called the Information
Logistics Process (ILP). Input to the ILP in our
study is the customer requirements. From these the
ILP communicates with the knowledge base to
produce a list of modules and parameters but also
marking failures, see Figure 1.
Figure 1: The Information Logistics Process (ILP)
The ILP processes are implemented with
different methods found in the computer science
area. Simple ILP processes are just straightforward
database solutions; other need real-time machinery
in order to deliver the information product at right
time to the right place. Real-time components can
handle time management and communication details
to facilitate the distribution of information to right
place in right time. (Frauenhofer Institute, 2006).
Still others need knowledge management. A
question is what can possibly be automated and what
will still be contained in the dialogue between users
of ILP and the ILP process. The ILP processes have
to handle knowledge, or more properly, expressions
of knowledge descriptions. This paper focuses on
the possible use of a knowledge base to partly
automate a process.
3 THE THREE PHASES OF THE
INFORMATION LOGISTICS
PROCESS
In Apelkrans and Håkansson (2005) we provided
rules for e-invoicing between enterprise systems as
an ILP process. In that case a company handled the
process as a third part Information Logistics
Provider (3ILP). A closer look at the ILP process
points out three different phases:
The first phase is for information supply (IS),
the information needed to trigger the next
information production phase. As an example
a customer sends invoice data to 3ILP and in
our study the input to ILP will be a customer’s
requirements presented with data and rules.
The second phase handles information
production (IP). In the 3ILP example IP will produce
correct information for the receiver’s ES and in the
ENTERPRISE SYSTEMS CONFIGURATION AS AN INFORMATION LOGISTICS PROCESS - A Study
213
study of this paper IP is searching for the correct ES
modules and the parameters that are affected. The
results are handed over to the last phase for
distribution.
The third phase is for information distribution (ID).
In the 3ILP case the desired receiving ES are
connected and information is delivered. In our work
the desired output is put together from the ILP
process, i.e. suggested modules and parameters are
presented to the user. These three phases are
presented in Figure 2.
Figure 2: The Information Logistics Process three phases.
The reason why ID is completed with a
possibility to use hard material can be illustrated in
the 3ILP case with the fact that some receivers of
invoice can’t take it electronic so we must produce a
paper variant and send by ordinary mail.
As mentioned earlier, all phases in ILP need to
handle knowledge descriptions in some way. Such
knowledge descriptions need to be specified from
case to case. A general question is: how should this
knowledge be elicited, stored and maintained?
Another question is: is it possible to automate
knowledge management in some way?
The IP phase is of course the main one of those
three and can be further divided into three phases.
Figure 3: A proposal for an IP architecture
One phase for creating information, another for
reproducing already created information, and finally
a phase leveraging the produced information to the
Information Distribution (ID) sub-process. Figure 3
below gives a suggestion of the architecture of the
Information Production (IP) sub-process.
A simple example of the IP sub phases is
handling a CD request containing a number of
tracks. The request needs be produced once and if
another customer wants the same CD, the request
simply has to be reproduced with little or no cost
and completed with information to be sent to the
right receiver.
In order to handle both information and
predefined rules for managing the information, i.e., a
kind of normative knowledge descriptions, the
architecture contains an Information Base (i.e.,
database) and a knowledge base. In the Information
Base one will find actual and/or stored information
and the knowledge base contains rules for their use.
In our study, most work (which is a matching rule
system) is done in the Information Creation sub-
process.
Our method contains an outer loop, which means
that ILP is called several times in order to refine the
module and parameter selections. The method also
contains an inner loop for comparing the customer’s
requirements with the contents of the system.
4 COMMON MODULES AND
PARAMETERS
An ES is a standard software package usually
packaged in a number of modules fitting different
application areas. As an example we present the
following list of modules:
General Ledger
Fixed Assets
Sales & Receivables
Relationship Management
Service Management
Purchase & Payables
Inventory
Manufacturing
Capital Requirements Planning
Human resources
These modules are possible to configure, i.e., set
certain parameters, to omit certain parts of a module
and possibly adding some extra software. The
behaviour of an ES module depends on the values
permitted to those parameters, which are decided by
the user. As an example of parameters, we give the
following list for the invoice function, see Table1:
ICEIS 2007 - International Conference on Enterprise Information Systems
214
Parameter Possible values
Costing Method Standard, FIFO LIFO
Specific Average
Price/Profit
Calculation
Profit = Price-cost, Price
= cost + profit
Sales Units: piece, 10, dozen, 100
Unit cost:
Number of
decimals
0,1,2,3
VAT calculations: %-age groups
Product type Material, part, complete
product
Table 1: Parameter types with possible values
As mentioned above, our proposed method has
an outer loop, a dialog between the user and the ILP
knowledge base. From user requirements, written in
rule form, this loop describes in more detail the
modules needed, the appropriate parameter list and
its parameter values. Finally the loop will come up
with a suggestion for software in the ES modules.
This software might be redundant since it is difficult
to find unused code in the system. Moreover, this
unused code might be needed later on. An open
question is still how to present the outcome to a
given ES.
The knowledge base in our ILP process contains
rule descriptions of every module there is in an ES.
In many cases a function in a module needs to be
detailed into sub-functions. These sub-functions are
also expressed by rules.
The inner loop of our method compares a
customer’s requirements rule description with those
in the knowledge base in order to find matching and
mismatching between them.
The ES modules are graphically described by
ARIS diagrams (Scheer et al, 2002) like the one
shown in Figure 5. Those descriptions are fairly
complicated and need to be structured in graphs with
sub-graphs. We call them functions with sub-
functions or processes with sub-processes. How to
handle them in a rule-based system is covered in
Section 5.
5 BUILDING RULES FROM
REQUIREMENTS
Customers have requirements on the Enterprise
Systems to adjust it to the organization. To meet a
customer’s specific requirement, the ES has to be
configured. Configuration aims at choosing
appropriate modules, refining the modules but also
setting their parameters to optimal values.
As mentioned before, the customer’s
requirements are originally presented in graphical
forms (using the ARIS diagrams suggested by
Scheer et al, 2002), which are designed as schedules.
In the schedules, the customers are specifying the
parts they need in a module that best suits their
business by developing, changing and extending the
schedules in the ES. We use these graphical
specifications to develop rules for the module.
Since these software packages can contain many
modules with thousands of parameters we have to
handle the requirements in an efficient way. Firstly,
a user has to study the ES Basic Terminology List
(for the list see e.g. Navision Attain 2002) in order
to avoid mismatches caused by the wrong
terminology. This is needed since the current system
does not detect the use of wrong terminology.
Secondly, the user must insert the parts of the
schedules of the processes used by the system and
make these as complete as possible. The parts used
in the system are triggers, functions, application
system types and organizational units. Each part has
its form and colour to distinguish them from one
another. The different parts are presented in the
figure below, Figure 4.
Figure 4: Different parts used in a schedule
The triggers are the events (e.g. “Start price
calculation”) handled by the system, which either
are inputs to functions or outcomes (“Result from
price calculation”) from functions. The functions, on
the other hand, are the actions provided by the
system (e.g. the actions needed to calculate the
correct price). These functions are some tasks to be
performed by the system. As mentioned above, some
of the functions have to be detailed in a sub-
function. The application system types can
symbolize different things but in our case they
describe a temporary storing of information. The
information is stored in a database to be used later
on in the graph. The organization units point out the
ENTERPRISE SYSTEMS CONFIGURATION AS AN INFORMATION LOGISTICS PROCESS - A Study
215
responsibilities of a certain action. Some of the
actions require some job to be carried out within the
company.
For expressing the parts in the software system,
we have chosen antecedent-consequent rules, i.e.,
production rules to represent the requirements. The
rules allow us to automate the configuration process
by using rule-based descriptions directly onto
customer’s requirement. The rules can be applied on
customer’s requirements because of the nature of the
requirements. These requirements use parts that can
be expressed as rules, like the logic operations,
triggers and actions (also called functions). The
logic operation, with “and” and “or”, can be applied
in rules by letting the “and“ be captured within the
rules and the “or” be captured by using several
different rules. Moreover, the triggers are applied as
facts and actions as rules.
The logic interpretation of the requirements is
the execution order of the rules. The order is the
relationships between the different rules. For
example, if one rule comprises another rule, both
have to be interpreted and completed before both are
usable in the conclusion. Moreover, if the rule
comprises two different rules, R2 and R3 in that
order, the rule R2 will be tested before rule R3. The
order those rules are applied in the rule becomes the
order of the execution.
The output from these rules, so called
conclusions, is a list of modules and parameters.
This list is used for configuration of the enterprise
system by comparing the list with the contents of the
system also expressed in rules. This list is checked
for correctness of the modules and parameters,
manually, and then for configuration the system’s
modules and parameters, automatically.
5.1 Example of Rules Built from Users’
Requirements
The customer’s requirements for the Enterprise
System are drawn as a schedule. This schedule
contains the triggers, functions, application system
types, organization units, logic operators and
parameters needed by the module to accomplish a
task. However, for the rules some of these parts are
omitted. The rules handle the triggers, functions,
logic operators and parameters. The organization
units are the units that take care of the product at the
company and do not need to be expressed in the
rules. Neither does the application system types need
to be expressed because it is fetches and sends data
to and from the database.
An example of building rules from the
customer’s requirements is using the schedule for
the invoice and production process, see Figure 5.
This process can be described with a couple of rules.
For example, applying rules on top of this schedule
can be done by the use of only one rule, but for the
quality of the knowledge base and following the
schedule, we use several rules.
Figure 5: An example of invoice and production process
(For details see appendix.)
There will be an overall rule that works as an all-
embracing rule, which invokes the facts and the
rules needed to reach a conclusion. In this example,
the overall rule starts by invoking the rule
Purchased order” and then calls for several
rules, i.e., “
Price calculation”, “Production
scheduling” and “Send invoice” to reach the
conclusion “
Invoice processed”. The
conclusion to the rule “
Purchased order” is
Purchased order received” and to the rule
Price calculation” “Price calculation
completed
”. We also have the conclusions
Production scheduling completed” and
Invoice Sent”.
The system uses both pre-stored facts, i.e., stored
in the database, and user-given facts, i.e., facts
inserted by the users during consultation with the
system. Hence, the facts, which correspond to the
triggers, are found by searching in the database. If
these facts are not found, the system asks the user. In
this example, the system is launched by the rule
“Process Invoice” which calls the rule
“Purchased order”. Then the latter rule invokes
the question “
Order to be created” which
answer or answers (if several answer alternatives)
became the user-given fact. Thus, the answer
becomes a fact that is initiating the consultation. The
fact “
Price calculation initiated”, on the
other hand, may be a pre-stored fact because the data
can be stored in the database of the system. Then,
this trigger only checks whether or not the session
ICEIS 2007 - International Conference on Enterprise Information Systems
216
has all data to be able to continue the consultation of
the system.
Besides invoking the fact “
Order to be
created”, the rule “Purchased order” call rule
Receive purchase order”, which can have a
schedule of its own that is the sub-functions
mentioned above. Thus, every function may use a
schedule like the one presented in Figure 5. This
schedule also needs to be executed before continuing
in the consultation.
Following the schedule in Figure 5, the rule
Price calculation” invokes the rule
Calculate price”, the fact “Price
calculation initiated
” and the rule “Issue
invoice
”. The rule “Production scheduling
invokes the rule “
Plan production”, fact
Production scheduling initiated” and the
rule”
Verify production plan”. Finally, the
rule “
Send invoice” is checked. All the rules in
this example are not specified, e.g., the last rule,
which either use a fact or a schedule. Even though it
is not pointed out in this paper, it illustrates the
complexity of rules that are produced even by small
examples.
The corresponding knowledge base, to the rules
presented above, is expressed in a kind of logic
syntax:
Rule “Process Invoice”->
Rule “Purchased order” and
Rule “Price calculation” and
Rule “Production scheduling” and
Rule “Send invoice”.
Rule “Purchased order”->
Fact “Order to be created” and
Rule “Receive purchase order”.
Rule “Price calculation”->
Rule “Calculate price” and
Fact “Price calculation initiated” and
Rule “Issue invoice”.
Rule “Production scheduling” ->
Rule “Plan production” and
Fact “Production scheduling initiated”
and
Rule “Verify production plan”.
Rule “Send invoice”.
Once these rules have been developed and
established into the system, they will become the
customer’s requirement rules.
5.1.1 Visualising the Developed Rules
Expressed in the syntax of a programming language
these rules together with the
relationships can be difficult to handle. To support
the users in understanding the rules, the rules and
relationships can be illustrated visually. As visual
tool, we use diagrams similar to the ones in UML
(Booch et al., 1999; Jacobsson et al., 1999). Some of
these diagrams have proved to be useful for rules
modeling the Information Logistic, as shown in
Apelkrans, and Håkansson, 2005, and modelling for
acquiring knowledge in Håkansson, 2001.
One might argue to use the schedules as presented in
Figure 5 but this does not illustrate the execution
order of the rules. Therefore, we will use a UML-
alike sequence diagram to present the rules
described above, see Figure 6.
Figure 6: A visual presentation of the rules in the
example of invoice and production process
In the figure the overall rule “
Process
Invoice
” is presented. It begins with following the
lifeline by invoking the related rules. In this case, the
execution is starting with “
Purchased order
illustrated as a square with broken lines. This rule
explores the fact “
Order to be created” and the
rule “
Receive purchase order”. After checking
these rules, the overall rule continues with the rule
Price calculation“ presented in a so-called
package. In this diagram, the package icon is used to
denote that there are contents in the rule and to see
it, the package has to be unfolded.
There are also two other rules in packages, the ”
Production scheduling“ rule and the “Send
invoice
” rule. At the end of the lifeline, the
conclusion is presented, i.e., for this schedule it is
Invoice processed”.
ENTERPRISE SYSTEMS CONFIGURATION AS AN INFORMATION LOGISTICS PROCESS - A Study
217
The functions, or actions, sometime have to be
detailed in a sub-function. These sub-functions are
also rules, which are hidden in the package. To
examine the package contents, the package is
unfolded. Unfolding the rule “
Price
calculation“ is to display the contents of that
rule, see Figure 7.
Figure 7: A visual presentation of the content of a package
Unfolding the rule package supports the
customer checking the contents. In this example, it is
found that one rule consists of two other rules
“Calculate price” and “Issue invoice”
and a fact Fact “Price calculation
initiated”. This folding/ unfolding feature is
necessary to make the rules comprehensible for the
customer. Too much information makes it hard to
get an overview of the rule-base whereas packages
hide important information.
6 MATCHING REQUIREMENT
RULES TO THE KNOWLEDGE
BASE RULES
The users build the schedules through drawing and it
is not certain that they will remember all the parts
that have to be in the schedule. For instance, in the
example presented above, Figure 5, the users may
have omitted some important parts, for instance, the
parameters.
The rules developed on the requirement schedule
have to be tested for viability and usability in the
system. The system must have a full collection of
full-fledge modules that contain all the parts and
parameters used in the system. These modules are
implemented as rules stored in the knowledge base.
These rules are used as templates to which the
developed rules, also called requirement rules, are
tested. For each part that is omitted in the
requirement rules, the system needs to ask whether
the users purposely omitted the parts or accidentally.
There might be another mismatch between the
rules in the knowledge base and requirement rules
do to the use of terminology used by the ES and
implemented in the rules in the knowledge base. The
system must find these mismatches, both from
checking the rules in the knowledge base as
presented above, but also from the requirement
rules. If parameters are omitted, the system asks for
answers on these. In the invoice case, it is to ask for
the costing method, price, profit calculation, sales
unit, unit costs, VAT and product type. Each of
these become facts in the rules. If a rule is wrong,
the user is asked to check it. The user either needs to
change the rule or reduce it. Moreover, if the rules
are in the wrong order, the user is asked to change
the order of the rules. It should be a possibility to
override these as well but then that enterprise system
may not be guaranteed to work properly.
If there are parts in the requirement rules that are
not present in the rules of the knowledge base, the
system asks the intension of these parts the users
have specified. The users are asked to check the rule
and its content but also check the terminology.
6.1 The Result of Applying Rules
The result of matching all the knowledge base rules
against the requirement rules is a list with the
modules and parameters (with chosen parameter
values) the users have requested. This list is
presented to the customer or user. The user should
carefully study the behaviour of the ES to be able to
suggest the right content of the modules. The user
might not be satisfied with the list and the modules
and/or the parameters need to be changed to undergo
the matching process again. This will be iterated
until the user is satisfied with the list. However, if
the user agrees to the list there are two options.
Either the list is directly applied to the ES or printed
as a report of what should be manually adjusted in
the system.
6.2 Verifying the Knowledge Base
The Enterprise system’s ability to expand the rule
base with the requirement rules for modules and
parameters depends on the ability to assure
correctness and completeness of the knowledge.
ICEIS 2007 - International Conference on Enterprise Information Systems
218
Therefore, the system should handle verification of
the knowledge base. Verification is the
demonstration of software consistency,
completeness and correctness at each stage.
However, in this work we only work with the
consistency.
Developing and updating the rule base involves,
among other things, checking the consistency. In
rule-based systems the rule base is consistent if for
each interpretation all facts are true (Beauvieux,
1990). Consistency checking includes testing
whether the system produces similar answers to
similar questions (Polat and Guvenir, 1993).
Consistency problems that can occur are conflicting,
redundant, subsumed and circular rules (Polat and
Guvenir, 1993). A conflicting rule is when it
contains a fact that is both true and false at the same
time. Redundancy occurs when several premises,
that are identical, are included in a rule. A subsumed
rule means that two rules produce the same result
but one is more restrictive than the other. Circular
rules is when a rule has dependencies to other rules
that prevent the rules to reach any conclusions.
Those rules have premises that use each other.
Depending on the schedules of the user, the system
must check so it will not run into verification
problem. This is an automatic test and should be run
for every schedule. If there is a cross-reference
between several schedules, these are tested together.
7 CONCLUSIONS
In this paper, we have presented a rule-based
description of the customer’s requirements on an
Enterprise System. Input to the system is the form of
requirement schedules and output is the modules and
parameters to be used by the Enterprise system. The
rules built from the requirements are matched to the
rules in the knowledge base. The missing data
between these rules will be asked for.
This study is restricted to a few examples of
handling an ILP process configuration of a customer
chosen ES. Therefore, there are a number of rules
and situations for several different ES that need to be
examined further. This might require adjustments of
the rule structures or the modules used in the
specific ES.
Future research will focus on managing the outer
loop, automating the process of transferring user
requirements given in graphical form to rule
systems. We will also build in the terminology for
the system. In the current version, the user has to
study the ES Basic Terminology List in order to
avoid mismatching by wrong terminology. In the
coming version, the system will present a list with
terminology used by the system from which the user
can use the right word. Since the number of terms is
vast, this list will be long so the user will either
choose from the list or write the words directly in
the interface of the system. If the user misspells the
word or uses wrong terminology, the system will
suggest the words that are commonly used in the
context for the modules.
Configuration costs a lot of money today, and
sometimes does not give a satisfying solution. On
the other hand, our proposal will also cost a lot of
money to fully implement the time-consuming effort
to build up the Enterprise systems Knowledge Base.
Each specific ES needs its own Knowledge Base
completed with rules for modules, actions and
parameters. Furthermore, a drawback with our
suggestion can be that graphical descriptions may
not familiar to ES users and therefore difficult to
use.
REFERENCES
Adam, F. and Sammon, D., 2004. The Enterprise Resource
Planning Decade: Lessons Learned and Issues for the
Future. • Idea Group Publishing.
Apelkrans, A. and Håkansson, A., 2005. Visual
knowledge modeling of an Information Logistics
Process - A case study. ICICKM-2005, 2nd
International Conference on Intellectual Capital,
Knowledge Management and Organisational Learning
Dubai, Förenta Arab Emiraten.
Beauvieux D. 1990. A general consistency (checking and
restoring) engine for Knowledge base. Proceedings
from ECAI-90. Pitman Publishing, p. 77-82.
Booch, G. Rumbaugh, J., and Jaconson, I. 1999. The
Unified Modeling Language User Guide. Addison
Wesley Longman, Inc.
Buck-Emden, R., 2000. The SAP/R3 Systems: An
introduction to ERP and business software technology.
Addison-Wesley
Frauenhofer Institute 2006
www.isst.fhg.de/englisch/download/34868_I-Log-4-
Seiter-engl-2.pdf, 10 december 2006.
Hedman, J. and Kalling, T., 2002. IT and Business
Models. Liber
Håkansson A., 2001. UML as an approach to Modelling
Knowledge in Rule-based Systems. ES2001 The
Twenty-first SGES International Conference on
Knowledge Based Systems and Applied Artificial
Intelligence. Peterhouse College, Cambridge, UK;
December 10th-12th, 2001.
Jacobson, I. Booch, G. and Rumbaugh, J. 1999. The
Unified Software Development Process. Addision
Wesley, USA.
Keller, G. & Teufel, T., 1998. SAP R/3. Process Oriented
Implementation, Addison-Wesley.
ENTERPRISE SYSTEMS CONFIGURATION AS AN INFORMATION LOGISTICS PROCESS - A Study
219
Navision Attain Terminology handbook, 2002
Olson, D. L, 2003. Manegerial issues of ERP
Systems.,McGraw-Hill
Polat F. and Guvenir H.A 1993. UVT: A Unification-
Based tool for Knowledge Base Verification. IEEE
Expert , June, No. 3.
Scheer, A-W., Abolhassan, F., Jost, W. and Kirchmer, M.
[eds] 2002. Business Process Excellence. ARIS in
Practice. Berlin: Springer Verlag.
Summer, M., 2005. Enterprise Resource Planning.
Prentice Hall.
APPENDIX
Figure 5 in full scale
ICEIS 2007 - International Conference on Enterprise Information Systems
220