Generating Process Entity Hierarchies from XPDL Process Models

Hyun Ahn

, Kyoungsook Kim

and Kwanghoon Pio Kim

Department of Computer Science, Kyonggi University, Korea, Republic of

Department of Computer Engineering, Kyunghee University, Korea, Republic of

Keywords:

Business Process Intelligence, Hierarchical Property, Process Entity Hierarchy, Hierarchy Generator, XPDL

2.0.

Abstract:

Business process intelligence enables us to discover a variety of deep insights about business process execu-

tion, and it provides a set of useful methods for related decision-making activities. The hierarchical infor-

mation that this paper focuses on is an important sort of information and it ought to be used in analyzing

hierarchical properties of business processes. In this paper, we present a useful hierarchy generator to make it

easier to perform analytics of hierarchical properties among business process entities. To this end, we deﬁne

an abstracted meta-model that represents hierarchical relations among entity types in XPDL process models.

According to the relational rules of the meta-model, a process entity hierarchy can be organized, analyzed,

and visualized.

1 INTRODUCTION

Business process management (BPM) technologies

and systems have been widely adopted in the various

industries where business processes are considered as

a critical factor in determining the enterprise value.

Accordingly, the importance of intelligent methods

for effectively analyzing business process increases to

enrich the managing capabilities of process-oriented

organizations.

Business process intelligence (Grigori et al.,

2004) and process mining (van der Aalst and Weijters,

2004) both are the highly active research ﬁelds with

the aim of analyzing business processes and discover-

ing valuable business process-related knowledge. In

this regard, a variety of approaches and methods are

proposed based on the fundamental aspects of the

business process, such as control-ﬂow, data-ﬂow, re-

source, and performance.

The hierarchical structure that this paper focuses

on is also an important kind of analytical property

pertaining to business process modeling and execu-

tion. The main purpose of the analysis of hierarchi-

cal information structure is to understand the struc-

tural characters of information elements and to eval-

uate each element by its proportion. Likewise, hier-

archical structures and analytical properties exist in

business processes and in the environment that man-

ages them (e.g., the explicit hierarchies among entity

types in business process models).

In this paper, we present a useful process entity

hierarchy generator to make it easier to perform ana-

lytics of hierarchical structures consisting of business

process entities in XPDL (XML Process Deﬁnition

Language) process models (WFMC, 2012). More

speciﬁcally, the generator is supported by the relation

rules based on the abstracted meta-model in the hier-

archy generation phase. Additionally, as a validation

step, we present a simple running example using the

sample models to generate, analyze, and visualize the

process entity hierarchy.

2 RELATED WORK

In this paper, we concentrate the hierarchical structure

and property implied the XPDL-based process mod-

els. In this regard, this section brieﬂy introduces the

previous works regarding the business process model

analysis.

Existing methods and techniques for business pro-

cess model analysis range from the fundamental prop-

erty analysis (van der Aalst and van Hee, 2004; Mu-

rata, 1989; Li et al., 2004) to the advanced and sophis-

ticated solutions (Dijkman et al., 2012; Khlif et al.,

2017; Awad, 2007; Kunze et al., 2011) to meet practi-

cal objectives and to achieve advantages in managing

business processes. Two prior works (van der Aalst

690

Ahn, H., Kim, K. and Kim, K.

Generating Process Entity Hierarchies from XPDL Process Models.

DOI: 10.5220/0006756106900695

In Proceedings of the 20th International Conference on Enterprise Information Systems (ICEIS 2018), pages 690-695

ISBN: 978-989-758-298-1

and van Hee, 2004; Murata, 1989) described the se-

ries of concepts and formal methods for fundamental

properties of the process model, such as reachability,

soundness, and liveness. In other hands, (Li et al.,

2004) presented the mathematical basis for measur-

ing the performance metrics including workload, and

turnaround time on the extended Petri nets.

While those prior works mentioned above con-

tributed to the foundation of modeling and verifying

business process models, more and more researchers

have been interested in developing methodologies to

efﬁciently manage a large collection of business pro-

cess models. (Awad, 2007) introduced the visual

language for expressing queries over a repository of

business process models. The author argues that this

language allows the reusability of business process

model without redundant efforts of modeling. Sim-

ilarly, (Kunze et al., 2011) presented the measure of

model similarity to search similar business process

models, which satisﬁes certain properties.

Although the other works are not included in this

paper due to the limitation of space, there is a volumi-

nous set of contributions which focus on business pro-

cess analysis. However, the hierarchical property of

business process has been considered as an insigniﬁ-

cant feature in the literature. In only a few studies, the

hierarchy concept has been used to express subpro-

cess (van der Aalst and van Hee, 2004) and resource

models (Du and Shan, 1999; Kim et al., 2005), but

neither of them addresses the analysis with the hierar-

chical structure composed of process entities.

3 AN ABSTRACTED

META-MODEL

In order to generate and analyze process entity hierar-

chies, we formerly devised the abstracted meta-model

(Ahn et al., 2016) from the meta-model deﬁned in the

XPDL 2.1 speciﬁcation (WFMC, 2012). This section

describes the details of the abstracted meta-model and

hierarchical relations among entity types.

3.1 The XPDL Standard

XPDL is the popular one of the standards for inter-

changing business process deﬁnitions between het-

erogenous BPM products. It was introduced by the

WfMC (Workﬂow Management Coalition) in 1998

and is currently standardized to the version 2.2. Fig-

ure 1 shows the meta-model of XPDL 2.1 speciﬁca-

tion representing primitive entity types and their rela-

tions. These entity types are broadly related to three

main aspects: Control-ﬂow, resource, and data.

Figure 1: The meta-model of the XPDL 2.1 standard.

3.2 The Abstracted Meta Model and

Hierarchical Relations

To facilitate the analytics of process entity hierarchy,

we analyzed the meta-model (shown in Figure 1) and

devised the abstracted meta-model representing the

hierarchical relations between entity types.

Package

Workflow

Process

Activity TransitionParticipant

Data Field

Pool

Lane

Application

Figure 2: Abstracted meta-model representing hierarchical

relations (composition and directed association) between

XPDL process entity types.

The abstracted meta-model includes essential en-

tity types which satisfy a hierarchical property and are

can be used as a construct to build a process entity hi-

erarchy. The deﬁnitions of entity types are described

as follows:

• Package: A package corresponds to a set of busi-

ness processes. In general, it means a grouping of

business processes in a speciﬁc business domain.

Therefore, a package is mapped to the root ele-

ment in an XPDL-formatted process model.

Generating Process Entity Hierarchies from XPDL Process Models

691

• Workﬂow Process: It refers to a business process.

In an XPDL ﬁle, a <WorkﬂowProcess> element

is a container for the set of its activities and their

temporal ordering.

• Activity: A typical activity represents a logical

unit of work in a business process. Also, the rout-

ing activities (e.g., OR split/join, AND split/join)

and events both are types of activity, but since

these types do not imply a hierarchical property or

related semantics we only take account of typical

activities in constructing a process entity hierar-

chy.

• Transition: Transitions are core elements of

control-ﬂow aspect in a workﬂow process. Each

transition connects from a predecessor activity to

a successor activity. In cases of transitions con-

nected from a disjunctive routing activity, a tran-

sition condition is attached to a corresponding

<Transition> element. Accordingly, the transi-

tion is activated only when the transition condi-

tion holds true.

• Pool: This entity type is the top-level container

for resources allocated in a business process. In

business process modeling systems, it is used to

graphically augment the view of the resources to

a business process diagram. According to the syn-

tax of XPDL, a <Pool> element includes a set of

<Lane> elements as child nodes.

• Lane: Lanes are used to describe responsibilities

of a certain set of activities in a business process.

Therefore, a lane is often relevant to a business

role (e.g., administrator and contract manager),

a system (e.g., a legacy system), a business unit

(e.g., department and project team).

• Participant: A participant is a resource capable of

performing an activity. It can be not only a human

resource but also a machine resource.

• Data Field: Data ﬁelds are created and consumed

within each business process execution. They

are used as variables to exchange intermediate re-

sults between activities or evaluate transition con-

ditions.

• Application: This entity type refers to a software

program or a service invoked from business pro-

cess enactments. They provide functionalities re-

quired to carry out tasks involved in a business

process.

Hierarchical relations the meta-model includes are

divided into two relation types: Composition and di-

rected association. A composition (depicted as –♦)

refers to an explicit hierarchical relationship between

a superior entity type and subordinate entity type.

These relationships are explicitly represented by hi-

erarchical XML elements in an XPDL ﬁle. For in-

stance, each <Package> element has a set of sub-

ordinate <WorkﬂowProcess> elements. A directed

association (depicted as →) refers to an association

relationship that is navigable in only one direction

and is interpreted as an implicit hierarchical relation-

ship, which is not explicitly represented in an XPDL

ﬁle, but it implies the semantic information related

the hierarchical properties of corresponding business

processes. For instance, an <Activity> element in-

cludes reference information about an identiﬁer of a

<Participant> element instead of explicit element in-

clusion.

4 GENERATING PROCESS

ENTITY HIERARCHIES

According to the abstracted meta-model, we imple-

ment a prototype system for generating a process en-

tity hierarchy that is based on the relational rules of

the meta-model. This section describes the hierarchy

generation steps with the implemented system. Addi-

tionally, we present a running example to validate the

feasibility of analytics of process entity hierarchies.

4.1 Process Entity Hierarchy Generator

A DCI

C DTD

CO D I

Activity

Participant

Data Field

Application

Package

Lane

Pool

Workflow Process

Transition

Figure 3: Tree representation of the hierarchical relations of

the abstracted meta-model.

The implemented generator is constrained by the rules

of hierarchical relation for the compliance with the

meta-model that we devised. Figure 3 shows the en-

tity tree representing hierarchical relations between

entity types. A hierarchy generation is implemented

by traversing the entity tree from the ﬁrst selected en-

tity type to a speciﬁc entity type that is the ﬁnal desti-

nation (except the root entity type). Additionally, all

ICEIS 2018 - 20th International Conference on Enterprise Information Systems

692

the entity types, excluding the leaf entity types, can be

the root entity type of a generated hierarchy since the

minimum required levels of hierarchy is two levels.

By using the graphical user interface provided by

the system, users can easily and interactively organize

a new hierarchy. The graphical navigation function-

ality offers users available options of entity types in

the phase of hierarchy generation. The simple exam-

ple of organizing a hierarchy is shown in Figure 4,

and its screen capture shows the situation of the hi-

erarchy generation. The hierarchy consists of three

layered entity types (package, pool, lane), and it is

currently incomplete and has only one available en-

tity type (participant) that the user can select.

Figure 4: Screen capture of organizing a hierarchy using the

graphical user interface.

After the user completes the hierarchy generation,

the hierarchy is automatically ﬂeshed out to the full

process entity hierarchy through the mapping of entity

information in XPDL models to the hierarchy. Figure

5 shows the generated process entity hierarchy. Con-

sequently, the user can perform analytics of hierarchi-

cal properties by applying analysis techniques to the

process entity hierarchy.

4.2 Running Example

To validate how useful the hierarchy generator is, we

demonstrate a simple running example using sample

XPDL process model.

The sample model contains one process package

(Commercial Banking) including two business pro-

cesses (Fund Transfer, Data-driven Fund Transfer),

and other entity information. From the view of re-

sources, the sample model also includes three pools,

eight lanes, and nine participants.

Figure 5: Generated process entity hierarchy

(Package–Pool–Lane–Participant).

The results of visual analytics executed by the im-

plemented system with the running example is shown

in Figure 6. At ﬁrst, entire entity information included

the input XPDL ﬁle is collected to the system through

the parsing phase (left side).

To analyze a hierarchy of resource en-

tity types, the process entity hierarchy

(Package–Pool–Lane–Participant) is generated

by the GUI-based hierarchy organizer and afterward

concretized with the corresponding collected entity

information (right side).

After the step of hierarchy generation is com-

pleted, the system visualizes the process entity hierar-

chy applied the pseudo-measures of work rate, which

means the portions of activity instances performed

by each participant. The visualization exploits the

treemap method (Bruls et al., 2000) and the color ﬁlter

(by each pool entity), and the node size represents the

work rate measures. Even though the visualization re-

sult is less effective to demonstrate the hierarchy and

related properties, we conﬁrmed that the feasibility of

the process entity hierarchy generation through this

running example.

5 CONCLUSION

To address analytical properties of the hierarchical

structure of business processes, we proposed the

XPDL-based process entity generator in this paper.

Generating Process Entity Hierarchies from XPDL Process Models

693

(a)

(b)

Figure 6: Visualizations of the analysis result of the generated process entity hierarchy applied the participants work rate

(activity instance) measures with two color coding methods: The pool-based color coding (a), the lane-based color coding (b).

The generator is governed by the rules of hierarchical

relations deﬁned in the abstracted meta-model. Con-

clusively, through the running example, we conﬁrmed

that the generator is fruitful to easily create a process

entity hierarchy from XPDL process models.

As a future work, we are planning to devise a com-

prehensive meta-model to ensure that the hierarchy

generator is compatible with other standards, includ-

ing BPMN, BPEL, and XES. As another issue, we

are interested in an effective visualization of the pro-

cess entity hierarchy. The visualization issue is not re-

garded as the main concern in this paper, but it is also

crucial to perform business process analytics. How-

ever, the presented visualization results are limited in

effectively displaying a process entity hierarchy and

its analysis result. To overcome this shortcoming, we

will be studying new visualizations that effectively

conveys the information or knowledge of hierarchical

properties of business processes.

ACKNOWLEDGEMENTS

This research was supported by Basic Science Re-

search Program through the National Research Foun-

dation of Korea (NRF) funded by the Ministry

of Science, ICT & Future Planning (Grant No.

2017R1A2B2010697). Also this research was par-

tially supported by the contents convergence software

research center at Kyonggi University funded by the

GRRC program (Grant No. 2017-0194) of Gyeonggi

Province, South Korea.

ICEIS 2018 - 20th International Conference on Enterprise Information Systems

694

REFERENCES

Ahn, H., Park, M., and Kim, K. P. (2016). Analyzing hier-

archical properties of entities in xpdl-based workﬂow

models. In Information Science and Technology, 2016.

11th Asia Paciﬁc International Conference on, pages

63–65.

Awad, A. (2007). Bpmn-q: A language to query business

processes. In EMISA, volume 119, pages 115–128.

Bruls, M., Huizing, K., and Van Wijk, J. J. (2000). Squari-

ﬁed treemaps. In VisSym, pages 33–42. Springer.

Dijkman, R. M., La Rosa, M., and Reijers, H. A.

(2012). Managing large collections of business pro-

cess models-current techniques and challenges. Com-

puters in Industry, 63(2):91–97.

Du, W. and Shan, M.-C. (1999). Enterprise workﬂow re-

source management. In Research Issues on Data En-

gineering: Information Technology for Virtual Enter-

prises, 1999. RIDE-VE’99. Proceedings., Ninth Inter-

national Workshop on, pages 108–115. IEEE.

Grigori, D., Casati, F., Castellanos, M., Dayal, U., Sayal,

M., and Shan, M.-C. (2004). Business process intelli-

gence. Computers in industry, 53(3):321–343.

Khlif, W., Khlif, W., Ben-Abdallah, H., Ben-Abdallah, H.,

Ben Ayed, N. E., and Ben Ayed, N. E. (2017). A

methodology for the semantic and structural restruc-

turing of bpmn models. Business Process Manage-

ment Journal, 23(1):16–46.

Kim, K.-H., Yoo, H.-J., and Kim, H.-S. (2005). A process-

driven e-learning content organization model. In Com-

puter and Information Science, 2005. Fourth Annual

ACIS International Conference on, pages 328–333.

IEEE.

Kunze, M., Weidlich, M., and Weske, M. (2011). Behav-

ioral similarity–a proper metric. Business Process

Management, pages 166–181.

Li, J., Fan, Y., and Zhou, M. (2004). Performance model-

ing and analysis of workﬂow. IEEE Transactions on

Systems, Man, and Cybernetics-Part A: Systems and

Humans, 34(2):229–242.

Murata, T. (1989). Petri nets: Properties, analysis and ap-

plications. Proceedings of the IEEE, 77(4):541–580.

van der Aalst, W. M. P. and van Hee, K. M. (2004). Work-

ﬂow management: models, methods, and systems.

MIT press.

van der Aalst, W. M. P. and Weijters, A. J. M. M. (2004).

Process mining: a research agenda. Computers in in-

dustry, 53(3):231–244.

WFMC (2012). Xml process deﬁnition language version

2.2.

Generating Process Entity Hierarchies from XPDL Process Models

695