A pattern‐based approach to extract REA value models from business process models

• Published date: 01 January 2017
• DOI: http://doi.org/10.1002/isaf.1402
• Date: 01 January 2017

RESEARCH ARTICLE

A pattern‐based approach to extract REA value models from

business process models

Anis Boubaker

1

|Abderrahmane Leshob

1

|Hafedh Mili

1

|Yasmine Charif

2

1

University of Quebec at Montreal (UCAM) –

LATECE Laboratory, Montreal, Quebec,

Canada

2

Xerox Research Center, Webster, New York,

USA

Correspondence

Anis Boubaker, Computer Sciences, Universite

du Quebec a Montreal, Montreal, Quebec,

Canada.

Email: anis@boubaker.ca

Summary

Business models are economic models that describe the rationale of why organizations create and

deliver value. These models focus on what organizations offer and why. Business process models

capture business activities and the ways in which they are accomplished (i.e. their coordination).

They explain who is involved in the activities, and how and when these activities should be per-

formed. This paper discusses the alignment between business models and business process

models. It proposes a novel systematic method for extracting a value chain (i.e. business model)

expressed in the Resources, Events, Agents (REA) ontology from a business process model

expressed in Business Process Model and Notation™. Our contribution is twofold: (1) from a the-

oretical standpoint we identified a set of structural and behavioural patterns that enable us to

infer the corresponding REA value chain; (2) from a pragmatic perspective, our approach can be

used to derive useful knowledge about the business process and serve as a starting point for busi-

ness analysis.

KEYWORDS

business processmodelling, business process patterns, REA, value modelling

1|INTRODUCTION

During the last decade, we have seen the emergence of the process‐

aware corporation paradigm as a replacement of the functional organi-

zation that has prevailed since the 18th century. The process‐aware

corporation seeks to foster agility by implementing business processes

that cross the borders of the traditional functional structure. Consider-

ing this new phenomenon and the resulting ubiquity of business pro-

cesses in today's corporations, business processes have been raised

as first‐class citizens by being considered corporate assets, and their

need for ongoing management has been recognized. In fact, business

process management (BPM) has been established as a dedicated

research area defined as the ‘methods, techniques, and tools to sup-

port the design, enactment, management, and analysis of operational

business processes’(Van der Aalst, ter Hofstede, & Weske, 2003).

Business processes are captured by business process models to

depict the set of actions that should be performed by a given business

process, in what order they should be performed and by whom. Lan-

guages such as Event‐driven Process Chain (Scheer, Thomas, & Adam,

2005) or Business Process Model and Notation™(BPMN; OMG, 2011)

define both a set of concepts and a graphical representation conven-

tions to model business processes. Therefore, the focus of business

process modelling is on the operational and dynamic aspects of the

process (i.e. How are we doing it? And who is doing it?). However

the business intent—expressing the whys of the activities we perform

—is not explicitly modelled and that is where business modelling comes

in handy.

A business model

1

is a conceptualization of the business process

using pure business terms such as economic resources, economic

agents and value‐adding activities and their relationships. Business

models help managers ‘to capture, understand, communicate, design,

analyze, and change the business logic of their firm’(Osterwalder,

Pigneur, & Tucci, 2005). It may take the form of a resource‐centric

view of the process, called a value chain, that permits answers to ques-

tions such as: How is the company using up its economic resources?

Why are the resources consumed or relinquished? There are three

main frameworks widely recognized in the literature: Resources,

Events, Agents (REA; McCarthy, 1982), e3value (Gordijn, 2002) and

business model ontology (Osterwalder, 2004).

1

The concept of business modelling is also called in the literature ‘value model-

ling’or ‘business case modelling’. However, ‘business modelling’seems to be

the generally agreed‐upon term.

Received: 15 January 2016 Revised: 24 July 2016 Accepted: 21 October 2016

DOI 10.1002/isaf.1402

Intell Sys Acc Fin Mgmt. 2017;24:29–48. Copyright © 2017 John Wiley & Sons, Ltd.wileyonlinelibrary.com/journal/isaf 29

In this study we propose a novel approach that aims to generate

a value chain (i.e. business model)—expressed using concepts from

the REA ontology (McCarthy, 1982)—from a business process model

expressed in BPMN. The problem is not trivial considering the large

conceptual distance between both views (Gordijn, Akkermans, &

van Vliet, 2000). Hence, our purpose is to infer business intentions

that are not explicit in the original business process model. We tackle

this by relying on a set of structural and behavioural patterns we have

identified through the analysis of a sample of business process

models. We propose a four‐step approach. In the first step we anno-

tate semi‐automatically the business process model to distinguish

economic resources. The second and third steps infer the REA con-

cepts from the process model in order to obtain a set of so‐called

REA economic processes. In the last step we connect the REA eco-

nomic processes in order to form a value chain of the business

process.

Although this work's focus is on the transformation from a BPMN

model to an REA value‐chain, the patterns we have identified render

explicit the relations between the two modelling abstractions notwith-

standing the particular notations used. Furthermore, we provide a for-

mal definition of our patterns so that they can be detected on any

business process modelling formalism given that the metamodel can

be transposed into our formal system.

In order to have early comments from the community on our

approach, a preliminary version of this work has been presented in

Boubaker, Cherif, Leshob, and Mili (2014). This paper extends the pre-

vious work on different aspects. First, in this paper we present our

second category of patterns: behavioural patterns. Second, we pro-

pose a formal definition of our structural patterns along with the tex-

tual presentation. This allows us to define our patterns unambiguously

in order to implement or analyse them. We also present in this paper

our algorithm to construct the resulting value chain. Finally, we

describe our implementation as a set of rules executed on top of a rule

engine.

This work has been conducted and is presented following the

design science research methodology as described by Peffers,

Tuunanen, Rothenberger, and Chatterjee (2007). In Section 2, we

present some background literature to briefly describe and distinguish

between business modelling and business process modelling. Section 3

will introduce the problem and the motivations of this work, as well

as a brief account on previous work that focused on relating the busi-

ness view with the business process view. The next four sections con-

stitute the design part of our research methodology. We will start by

giving a high‐level overview of our approach in Section 4, while

Sections 5 and 6 will focus on the inference of REA concepts and

REA processes based on five structural patterns and four behavioural

patterns that we will introduce. In Section 7 we describe our algo-

rithm to generate the value chain from the inferred REA economic

processes. To demonstrate the systematic application of our

approach, we implemented a tool as a rule‐based application. We pro-

vide an overview of our implementation in Section 8. The evaluation

of our approach will be performed empirically and we will describe

our experiment and analyse our results in Section 9. Finally, we will

conclude in Section 10 and raise some outstanding issues and future

work.

2|BACKGROUND LITERATURE

We distinguish in this paper between two modelling perspectives: (1)

business process modelling and (2) business modelling. The latter

describes the business rationale of the process (i.e. the why), whereas

the former takes an operational standpoint (the how, whom and when).

In the following we describe each of them in turn.

2.1 |Business process modelling: The BPMN

language

Companies design their business processes in order to express how to

‘do things’efficiently and effectively (Laguna & Marklund, 2013). In

order to document and communicate such designs, business processes

are modelled using a modelling language that provides a set of con-

cepts along with their syntax and semantics to express these designs;

usually as a set of orchestrated activities and communications between

participants. Apart from communication and documentation purposes,

business process models may be used to automate the process or to

analyse it either structurally (e.g. free of dead and live locks) (Palma,

Moha, & Gueheneuc, 2013; Van der Aalst, 2003) or semantically (e.g.

process re‐engineering). A business process model looks at the

dynamic and operational aspects of the business process.

As we can see from the survey paper by Mili et al. (2010), in recent

decades we have witnessed a profusion of business process modelling

languages, each inclined towards one or more modelling objectives (i.e.

documentation, execution or analysis). One of these languages, BPMN,

has received wide interest by researchers and practitioners alike, and

its second version was standardized by the Object Management Group

in 2011 (OMG, 2011). The BPMN standard provides (1) a modelling

notation and its semantics, (2) a serialization convention for both the

model and its graphical representation as XML schemas and (3) an exe-

cution semantics of the models.

Typically, a business process expressed in BPMN depicts a set of

activities, events and gateways (in BPMN terms: FlowElements)

ordered by a control flow relationship. The activities may require

inputs or produce outputs modelled as objects (ItemAwareElements)

that could be either informational or physical. The set of activities, con-

trol flow and item‐aware elements form a business process model. The

BPMN language also enables the modeller to design the interactions

between two or more Participants either as a Collaboration or Chore-

ography. The choreography is a message‐centred view of the multi‐

participant process, while the collaboration (the focus of this paper)

has a flow‐oriented view that integrates communications as part of

the flow. In a collaboration, each participant has its business process

enclosed within a Pool and a pool might be divided into swim lanes

to reflect a participant's internal organization.

BPMN permits the modeller to design a business process from dif-

ferent abstraction levels. Weske (2012) recognizes three modelling

levels: (1) the organizational level exhibiting coarse‐grained activities;

(2) the operational level, which covers more granular process activities

and their logical orderings from a business perspective; and (3) the

implementation level, in which a business process encompasses all

necessary details needed to be executed on an execution engine. In

30 BOUBAKER ET AL.