A Measurement-oriented Modelling Approach

A Step Forward

Giulio D’Emilia, Gaetanino Paolone, Emanuela Natale, Antonella Gaspari and Denis Del Villano

Department of Industrial and Information Engineering and of Economics, University of L’Aquila, L’Aquila, Italy

Keywords: Use Case, Business Modelling, System Modelling, UML, Measurement, Uncertainty, Energy.

Abstract: Measurements represent a fundamental component of Enterprise Information Systems and they play a key

role in organizations. Their own languages, concepts and techniques, concerning how to approach and solve

problems in modern industrial scenarios, inevitably characterize these two disciplines. This is why the

question we posed is to get a methodology that allows us to analyse, model and implement software

subsystems able to render really usable information concerning measurements, keeping their informative

peculiarities unchanged. The final goal of our research is to define a Use Case-based methodological

proposal for modeling the informative content of measurements and their usage that starts from the business

model of an enterprise subsystem and achieves a software model able to satisfy the users' needs.

1 INTRODUCTION

It is generally acknowledged that designing and

developing software systems is becoming

increasingly complex. Fortunately, there are

methodologies and tools (Sukaviriya et al., 2009) to

tackle this demanding and, sometimes critical,

challenge. For example, the methodology proposed

in (Paolone et al., 2008a; 2008b; 2009) promotes the

iterative and incremental development of complex

software systems using a methodological framework

that supports model-driven engineering. Such a

methodology is inspired to the Rational Unified

Process (RUP) (Kruchten, 2003) and it poses Use

Cases (UC) at the centre of the modelling (UML,

2012).

Nowadays measurements, i.e. quantitative

information from measured quantities, represent

more and more a fundamental component of

Enterprise Information Systems (EIS) and they play

a key role in organizations. While the automation of

decision-making processes based on measurements

appears to be a great opportunity, on the other hand

difficulties are presumable. There is the possibility

of having a large amount of data coming from

measurements to be integrated into the Business

Information System that have their primary language

and who are not always well spread throughout all

departments of the business organizations. Moreover

the source of information has an extremely wide

variability, in the measuring system implementing

methods and in the quality of measurements. There

are concepts related to the variability that may give

rise to content's smokiness and, then,

computerization may be a useful solution. Another

difficulty is that the operating conditions can have,

from one case to another, completely different

characteristics and connotations. This is why the

question we posed is to get a methodology that

allows us to analyse, model and implement software

subsystems able to render really usable information

concerning measurements, keeping their informative

peculiarities unchanged. Please note that in literature

there are very few examples that can be supportive

(Wen Bilong et al., 2009). Studies aiming to

compare foundations of measurement theory to

software measurement (Carbone et al., 2008) do not

appear, in fact, closer to these goals.

For an IT project to be successful, it must be as

close as possible to business reality, in such a way

that corporate users can find in the application (Zhao

et al., 2007) the same modus operandi of their own

function: each actor plays a set of UCs within the

organization and does so regardless of automation.

Today, UCs are at the core of modelling and

developing software applications (Zelinka, Vrani´,

2009) (Duan, 2009) (Sukaviriya et al., 2009). The

methodology appeared in (Paolone et al., 2009) is an

137

D’Emilia G., Paolone G., Natale E., Gaspari A. and Del Villano D..

A Measurement-oriented Modelling Approach - A Step Forward.

DOI: 10.5220/0005090601370142

In Proceedings of the 9th International Conference on Software Engineering and Applications (ICSOFT-EA-2014), pages 137-142

ISBN: 978-989-758-036-9

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

instance of the proposal that empower to manage

such a complexity through a layer of classes

dedicated to UC automation. Their methodology

examines the system behavioural aspects through a

top-down process (such an approach is

commonplace amidst software development

methodologies), and then proceeds by means of

stepwise refinements of the initial business model.

The final goal of our research is to define a

methodological proposal for modeling the

measurements and their usage that starts from the

business model of an enterprise subsystem and

achieves a software model able to satisfy the users'

needs (i.e., that fully adheres to business processes).

In line with this goal, the present contribution calls

into question the convenience of using a top-down

approach in business modeling, system modeling,

design and implementation of a software system able

to make available the expected information, arising

from the measurements, to the management.

The next step (started with this position paper)

adapts the approach proposed in (Paolone et al.,

2008a, 2008b, 2009), transforming it in such a way

that you can understand and design software

application for the analysis of measurements starting

from business system requirements. In summary,

what we want to do is to extract UCs from the EIS

and bring them into the computerized system (from

Business Modeling to System Modeling) also in

relation to the measurements to be carried out in any

enterprise area, whether they are related to the

production, power consumption or all other forms of

detection.

The paper is organized as follows. Section 2

recalls essential elements of the methodology

appeared in (Paolone et al., 2008a, 2008b, 2009)

needed for understanding this work. Section 3

outlines essential characteristics of an EIS'

subsystem dedicated to metering and its peculiarities

in decision-making, regardless of the usage of

computer. Section 4 starts the discussion about a

possible transformation of the methodological

process recalled in Section 2, which can lay a solid

foundation for pursuing the aforementioned ultimate

goal. Brief conclusions end the paper.

2 THE METHODOLOGY

The methodology introduced in (Paolone et al.,

2008a; 2008b; 2009) allows to represent in detail

two models: the business and the system model. Use

case modeling and realization are the most important

aspects of the methodology. The proposal is

centered around four distinct layers (Figure 1) with

an iterative and incremental approach that leads to

the realization of a Business Use Case (BUC) into

the software application through stepwise

refinements. The first two layers of UC analysis are

placed in the business modeling context: their

objective is to get a complete representation of the

given business reality. The next two layers are

instead placed in the system modeling context with

the objective of representing the software system.

More in detail, they say that the first layer concerns

BUCs analysis, which are then specialized by

Business Use Case Realizations (BUCR) in the

second layer. Afterwards, a trace operation is used

to define the system UCs (third layer), which are

then specialized by Use Case Realizations (UCR)

(fourth layer). The latter ones can be implemented

by Object Oriented classes.

Figure 1: A sketch of the methodological layers.

Next, we describe the methodology in detail

through a brief example referring to a real-life

document management project for a bank, where

every layer contains a type of UML diagram.

Figure 2: The BUC diagram (1

layer).

ICSOFT-EA2014-9thInternationalConferenceonSoftwareEngineeringandApplications

138

This example may be useful because, as we will

show hereafter, what is being developed for a bank,

which is a typical management case, can be applied

to any industrial scenario.

Figure 2 shows a fragment of the BUC diagram,

placed in the first layer of Figure 1.

The example shows how BUCs are used to

express an actor/system interaction. For each BUC,

we define the related BUCRs. Referring to the BUC

Documental Management, Figure 3 proposes six

BUCRs.

Figure 3: The BUC realize diagram (2

layer).

After the business modelling phase, we analyse

the part of the system that will be automated. The

trace operation can introduce many system UCs for

a single BUCR. For example, in

Documental

Management, the document acquisition can be done

by the Bank, but also by Suppliers (see Figure 4).

The output of the trace operation produces the

system UCs in the third layer of Figure 1.

Figure 4: The use case trace diagram (3

layer).

In the last phase of the subsystem behavioural

analysis, we must identify at least one system UCR

for each system UC. In this phase we also introduce

some technological UCRs, such as

LinkFile. For

the sake of brevity, we don’t present an example of

system UCR diagram, but it should be

straightforward to understand that this operation

introduces a further refinement of the subsystem.

The current methodology has a strong industrial

impact because it has been repeatedly applied in real

projects reaching good results and its adoption has

brought benefits both in terms of the engineering

aspects of design and development time (Paolone et

al., 2008a). Moreover, the methodology enables to

build software systems with the help of a an existing

Java-based framework that implements a Java class

for each UCR and permits to speeds up software

development.

In conclusion, it is possible to reaffirm that the

methodological process is UC-driven, since the UC

artefact exists both in the business model and system

model, although it is represented by different

stereotypes, and is also exported to code.

3 THE MEASUREMENT

VIEWPOINT

Decision making requires both information and

knowledge. Information (or its absence) is central to

decision making (Beretta et al., 2012). In other

circumstances the theory of measurement has alredy

demonstrated to favor the ability to enter in the

actual reality of the processes of interest (D’Emilia

et al., 2014a). Therefore, information deriving from

measurement data may play a key role in business

decision making. In business management, it is

important that decision is supported by appropriate

tools, having the function to give the possibility to

minimize the risk of underestimate and of errors to

the called person to make a complex decision. In this

sense, measurement uncertainty offers a

considerable aid to quantify that risk, because it

refers to the concept of the information reliability

level (level of confidence).

In fact, if the data are accurate, i.e. closer to the

"true quantity value" of the measured quantity, then

they can be processed effectively creating an

informative base with the following features:

 shared, i.e. integrated within business

informative systems set in the specific

industrial situation;

 transparent, i.e. objective and incontestable

from the team members who participate to the

decisional process;

 significant, i.e. consistent from the data

quality viewpoint,

 aware, involving, in other words, an

indication about the risk assumed by the

decision maker, with reference to different

alternative choices.

In that context, aiming the fulfillment of these

features, the attention must be paid to several

challenging aspects for both information systems

Document Acquisition

(from Business Use-Case Model)

Internal Document Acquisition

Document From Supplier

<<Trace>>

AMeasurement-orientedModellingApproach-AStepForward

139

and metrology disciplines. Without limiting the

general nature of the foregoing, an interesting area

of use of a decision-making strategy based on

measurement uncertainty of data coming from the

field, is referred to an energy case of optimization.

In particular, with reference to an industry operating

in the aeronautical sector, simple measurements

allowed us to validate a predictive model of energy

consumption, to be used for the definition of a cost

effective strategy for energy saving (D’Emilia et al.,

2014b).

In this context, the decision-making strategy

provides for the possibility of having a management

tool that, for example, is able to return alternately:

 the correspondence between a budget of

improvement (I) and the target (t) that can be

guaranteed, in front of a predetermined level

of confidence (k) or risk deemed acceptable

by the decision maker;

 the relation between a variable and adjustable

improvement investment I



and the probability

p(k') that a target set as t’ is achieved.

In fact, in order to ensure, for the same

investment I, and with a given level of confidence

p(k) to achieve a given objective t, it is necessary

that the model gives the value  as a solution,

which is related to the target of a quantity t exactly

equal to the measurement uncertainty of the model,

U(m), according to the following logical

implication:

p



t, I



%p





→mtU

with:







k∗u

where:

p(t, I)%: probability of reaching the target t, with

the investment I;

 : degrees of freedom;

k, k': coverage factors (with ∞);

p(k): probability (confidence level) associated with

the model with the coverage factor k chosen;

t, t’: target fixed or variable depending on

investment;

m: indication of the consumption model validated,

i.e. provided of its uncertainty, m = f (I);

m : indication of the model that is in new condition

after the fixed investment I;

m : indication of the model corresponding to the

realization of the investment variable 



;

u (m): standard uncertainty of the model;

U (m): expanded uncertainty of the model.

Furthermore it is possible to study the

relationship, p







f′I



, between probability p(k')

to reach the target and the required investment I



. In

fact, in front of an investment I



the model will return

an indication mfI



 corresponding to a reduction

in consumption plausibly less ambitious (i.e. m

m ), being: k’≠k.

4 THE APPROACH WE LOOK AT

Designing a large enterprise software application is a

complex and articulated process since it represents

the company automation. Particularly critical

appears the identification of the UCs that illustrate

the interaction modes of the end-users with the

system according to the usual business workflows. It

is important to emphasize that the usage of a

methodology, in the context of software engineering,

has a fundamental importance to dominate the

complexity of computerized solution.

As described in previous sections, measurements

represent a key element in decision making. BUCs

and BUCRs detection is a critical factor for the

success of software applications which aim to be

strategic for the business management and that are

inspired by measurements. As a first step towards

the definition of a methodology for the analysis and

design of software for decision making that is based

on measurements, we apply the methodology

mentioned in Section 2 to a real case. The case study

is referred to an energy case of optimization within

an avionics components’ enterprise: the main goal is

to reach the energy consumption optimization.

In the proposed approach, the business modeling

activity starts, in close collaboration with the

enterprise top management, from the detection of

Organization Units involved in the IT project and

then proceeds discovering their Business Systems

(BS) and their Business Goals (BG). Four BSs were

detected and analyzed: in the example discussed

hereinafter we focus on one of them, the BS

EnergyManagementArea, involved in reaching the

BG named

EnergyConsumptionEfficiency.

Inside every BS we identify Business Actors,

BUCs and BUCRs, using the construct BUC to

represent a single interaction mode between actors

and the system and the construct BUCR to represent

how business workers, business entities, and

business events collaborate to perform a particular

BUC (Johnston, 2004).

After a careful analysis of the Company, paying

particular attention to information flow inside the BS

ICSOFT-EA2014-9thInternationalConferenceonSoftwareEngineeringandApplications

140

EnergyManagementArea

, we identified several

BUCs. In presenting our proposal, particularly

interesting are the BUCs performed by the Business

Worker

Energy Manager

, whose decisions are

closely related to the measurements made on the

field. Among those several BUCs, the most complex

(from the knowledge-intensive point of view) is

ConsumptionTargetManagement

, realized by 3

BUCRs (Figure 5).

Figure 5: Part of the case study BUC realize diagram.

To better understand the logic flow and

document knowledge aspects involved in

knowledge-intensive BUCRs, we widely use

Business Activity Diagrams (BAD) (where a

Business Activity (BA) denotes an elementary

business operation or a knowledge-intensive task)

and a strong narrative description. The ability of

UML BADs to effectively describe complex

business processes (Russel et al., 2006) allows us to

depict the inference process that permits the

Business Actor to take a complex decision. A

complex BA (that is an activity representing a

number of intricate atomic tasks) may be depicted at

different grain-size levels through the use of several

BADs.

Figure 6: Part of the case study BA diagram.

For example, the BUCR

Target-Model

Comparison

– representing the concepts expressed

in Section 3 – was depicted using the BAD in Figure

6 (which is only a part of a larger diagram because

of space limits) and also widely documented through

a narrative specification.

During the execution of the business modeling

discipline, as provided for by theory, the main

Business Entities (BE) (representing a significant

and persistent piece of information that is

manipulated by Business Actors and Business

Workers (Johnston, 2004)) were also identified and

modelled.

Specific attention was paid to documenting

classes of measurement-intensive business objects,

i.e. those BEs strongly related to measurements. In

their modeling, close attention was paid to maintain

the peculiarities of measurement unchanged and

well-marked, in order to grant a key role in business

decision making to information deriving from

measurement data. Figure 7 shows a portion of the

BEs diagram

Figure 7: Case study Business Entities.

After the Business Analysis, a trace operation

was performed: according to the methodology, we

identified the BUCRs to be computerized and we

traced them into System UCs (Figure 8).

Figure 8: Part of the case study UC trace diagram.

In the same way, a trace operation was

performed only on BEs needed for the

computerization.

In the last phase of the subsystem behavioural

analysis, we identified UCRs for each system UC.

Each UCR was diagrammatically depicted in terms

of scenarios: for the sake of brevity we will not

present a figure of system UCR diagram.

AMeasurement-orientedModellingApproach-AStepForward

141

5 CONCLUSIONS

At the end of the case study’s modelling process we

believe the proposed approach produces a good

representation of the EIS to be computerized and a

concrete image of subsystems to be automated. It is

important to remark we achieved this firm belief in

close collaboration with several stakeholders

involved in various aspects into the project, mainly

measurements experts, decision makers, IT-business

analysts and software engineers. Therefore, in our

opinion, the usage of this methodological approach,

broadly integrated with the usage of BADs (mainly

to represents business decision-making patterns)

permits to improve the communications’ quality

between the various stakeholders involved in

modelling and designing a measurement-intensive

software system. Starting from the business

modeling activity, the increase in the information’s

quality may help to reach a more effective system

analysis and, at the end of the process, to build a

software system as close as possible to business

reality and fully able to reveal its decision-making

patterns.

Finally, we believe this approach may become a

first step in reducing the informative gap

(concerning the correct usage and interpretation of

measurements) between business management,

software engineers and measurement experts, giving

some preliminary solutions deriving from the fact

that in the best of our knowledge, measurements are

not correctly used into automated decision making

processes as often the typical concepts of

measurement (uncertainty, level of confidence, ...)

are lost while being processed and made accessible

to end-users. A change is needed in the usage of

measurements in decision making processes

modelling and computerization and the proposed

top-down approach may be a first step in this

change. Of course, to completely clarify how

measurements need to be correctly used and

interpreted within an automated decision-making

process, requires that many aspects are further

studied with reference to the business modelling, to

the type of approach (top-down, bottom-up, mixed),

to the procedures of in field transfer of the results,

etc.)

REFERENCES

Beretta, F., De Carlo, F., Introna V., Saccardi D., 2012.

Progettare e gestire l’efficienza energetica. McGraw-

Hill.

Carbone, P., Buglione, L., Mari, L., Petri, D., 2008. A

Comparison Between Foundations of Metrology and

Software Measurement. IEEE Transactions on

Instrumentation and Measurement, 235-241.

D’Emilia, G., Di Rosso, G., Gaspari, A., Massimo, A.,

2014a, Metrological interpretation of a six sigma

action for improving on line optical measurement of

turbocharger dimensions in the automotive industry, to

appear on Proceedings of the Institution of Mechanical

Engineering: Part D, Jnl. of Automotive Engineering

D’Emilia G., Gaspari A., Natale E., 2014b. Uncertainty

evaluation of energy flow in industrial applications as

a key factor in setting improvement actions. Proposed

for publication to Applied Energy.

Duan, J., 2009. An approach for modelling business

application using refined use case. ISECS

International Colloquium on Computing,

Communication, Control, and Management (2009).

Johnston S., 2004. Rational UML Profile for business

modelling. IBM Rational (www.ibm.com).

Kruchten, P., 2003. Rational Unified Process, An

Introduction (2

Edition). UK, Addison Wesley.

Paolone, G., Clementini, E., Liguori, G., 2008a. A

methodology for building enterprise Web 2.0

Applications. The Modern Information Technology in

the Innovation Processes of the Industrial Enterprises

Prague Czech Republic, 12-14 November 2008.

Paolone, G., Clementini, E., Liguori, G., 2008b. Design

and Development of web 2.0 Applications. ITAIS 2008

Paris France, 13-14 December 2008.

Paolone, G., Clementini, E., Liguori, G., Cestra, G., 2009.

Web 2.0 Applications: model-driven tools and design.

ITAIS 2009 Costa Smeralda (Italy) October 2-3, 2009.

Russell, N., van der Aalst, W.M.P., ter Hofstede, A.H.M.,

Wohed, P., 2006. On the suitability of UML 2.0

activity diagrams for business process modeling. 3rd

Asia-Pacific Conf. on Conceptual modeling.

Conferences in Research and Practice in Information

Technology, Vol. 53. M. Stumptner, S. Hartmann and

Y. Kiyoki, Eds.

Sukaviriya, N., Mani, S., Vibha Sinha, V., 2009.

Reflection of a Year Long Model-Driven Business and

UI Modelling Development. INTERACT 2009, Part II,

LNCS 5727, pp. 749–762.

UML, Unified Modeling Language, 2012. Version 2.4.1,

http://www.uml.org/

Zelinka, L., Vrani´, V., 2009. A Configurable UML Based

Use Case Modeling Metamodel. First IEEE Eastern

European Conference on the Engineering of Computer

Based Systems.

Wen B., Zhang L., 2009. Mapping Enterprise Process

Measure into Information Model. First International

Workshop on Education and Computer Science, pp.

612-615.

Zhao, X., Zou, Y., Hawkins, J., Madapusi, B., 2007. A

Business Process Driven Approach for Generating E-

Commerce User Interfaces. Model Conference 2007

Nashville TN September 30 – October 5 2007, pp.

256-270.

ICSOFT-EA2014-9thInternationalConferenceonSoftwareEngineeringandApplications

142