Towards an intelligent workflow designer based on the reuse of workflow patterns [original]

Towards an Intelligent Workflow Designer based on the

Reuse of Workflow Patterns

Cirano Iochpe, Carolina Chiao

Guillermo Hess

Gleison Nascimento

Federal University of Rio Grande do Sul

Institute of Informatics

Av. Bento Gonçalves, 9500

91502-970, Porto Alegre, Brazil

+55 (51) 33166820

{ciochpe, cchiao,

hess,gsnascimento}@inf.ufrgs.br

Lucinéia Thom

Ulm University

Institute of Databases and

Information Systems

Oberer Eselsberg

Geb. O27, Raum 515, 89069

+49 (073) 15024137

lucineia.thom@uni.ulm.de

Manfred Reichert

Faculty of Eletrical Engineering

mathematics and Computer

Science

University of Twente

7500 AE Enschede

The Netherlands

+31 (53) 489 3705

m.u.reichert@cs.utwente.nl

ABSTRACT

In order to perform process-aware information systems we need

sophisticated methods and concepts for designing and modeling

processes. Recently, research on workflow patterns has emerged

in order to increase the reuse of recurring workflow structures.

However, current workflow modeling tools do not provide

functionalities that enable users to define, query, and reuse

workflow patterns properly. In this paper we gather a suite for

both process modeling and normalization based on workflow

patterns reuse. This suite must be used in the extension of some

workflow design tool (e.g., Intalio, Event Process Chain – EPC).

The suite comprises components for the designing of processes

from both legacy systems and user process.

Keywords

Workflow Patterns reuse, Business Process, Legacy Systems,

Process Design, Ontology.

1. INTRODUCTION

For several reasons companies are developing a growing

interest in improving the efficiency and quality of their

internal business processes and in optimizing their

interactions with customers and business partners. During

the last years we have seen an increasing adoption of

business process management tools by enterprises as well as

emerging standards for business process specification in

order to meet these goals. Respective technologies (e.g.,

workflow management systems) enable the definition, exe-

cution, and monitoring of the operational processes of an

enterprise. In connection with Web service technology, in

addition, the benefits of business process management from

within a single enterprise can be transferred to cross-

organizational business processes as well.

Business Processes and respective workflow models

frequently include a variety of fragments (or recurrent

business functions) which can be understood as self-

contained activity blocks with a specific and well-defined

semantics [1], [2], [3]. As an example, let’s consider the

evaluation process for price adjustment as depicted in

Figure 1. This process includes activities with the following

partial order: (a) verify if it is a shopping order or not; (b)

evaluate request of price adjustment; (c) notify managers

about conclusion of evaluation; (d) notify managers about

automatic approval. Altogether this process comprises

fragments that presents generic semantics which can be

described as patterns such as decision (activity a),

notification (activities c and d), and task execution request

(activity b). In this paper, we are dealing with the question

of how the modeling of processes that include recurrent

business functions like notification in Figure 1 can be

supported appropriately by a tool. Moreover, how the

extraction of business process from legacy systems could be

accomplished when supported by such patterns.

Figure 1. Evaluation process for price adjustment

So far, several workflow patterns have been suggested for

representing control flow [5], resources [6], data [7] interaction

[8] and exception handling [9]. Yet, these pattern sets have in

common that they are relevant for the implementation of a

workflow system and the definition of process modeling

languages, but they provide only a partial answer to the question

of what business functions a modeler has to consider repeatedly in

various process models. Usually, such process fragments [10],

[11], [12], [13], [8] are re-designed for practically every workflow

application. Such procedure can be considered as inefficient, and

thus undesirable from a maintenance perspective. While there is

some research reported on how metadata can be organized to

manage large-scale modeling project (see [14]), we are not aware

of any work evidencing the existence of recurrent patterns in real

workflow applications as well as their necessity and completeness

for the business and workflow process modeling. Besides that,

contemporary workflow modeling tools do not provide

functionalities that enable users to define, query, and reuse such

patterns in a proper way.

Related to these problems we proposed a set of seven workflow

patterns in an early work [1]. Each pattern represents a recurrent

business function (such as the ones showed in Figure 1)

frequently found in business processes. In this paper we introduce

a suite for normalizing and modeling of business processes based

on the reuse of workflow patterns. By normalization we mean the

definition of a standard description form to which the business

processes are translated, i.e., a canonical format for describing

workflows.

This suite (the so called “Intelligent Workflow Designer”) must be

used in the extension of some workflow design tool (e.g., Intalio,

Event Process Chain – EPC), and is intended to provide a number

of functionalities, such as: (1) the extraction of business processes

from legacy systems and their normalization, correctness checking

in a formal notation and translation into a standard notation; (2)

support for process design, by suggesting to the modeler patterns

to be combined to the one he/she modeled and; (3) construction of

a knowledge base for storage and retrieval of workflow patterns.

Against this background, the outline of this paper is organized as

follows: Section 2 gives an overview of the workflow patterns that

we identified in prior researches. In particular, we discuss the

approval, unidirectional performative and the notification pattern

as three examples. Section 3 gathers a suite for the extension of

some modeling tool that aims to support the reuse of these

patterns. We describe each component of the suite which

considers process modeling from legacy systems or from user

processes (e.g., design from scratch). Finally, Section 4 concludes

the paper and gives an outlook on future research.

2. WORKFLOW PATTERNS

In the context of this paper we use the term workflow pattern to

refer to the description of a recurrent business function frequently

found in business processes (e.g., notification, decision,

approval). We derived a set of 7 patterns from an extensive study

based on the literature. Examples of patterns are document

approval, question-answering, unidirectional and bi-directional

performative, information, notification and decision patterns.

Details on these patterns as well as a classification of them are

reported in [2] and [3].

It is out of the scope of this paper to detail the semantics of all

these patterns. It is important to note that through the mining of

190 real workflow processes we measured the occurrence

frequency of each of the workflow patterns in the set of workflow

processes analyzed. In general words, the main results of the

mining can be summarized as follows:

–

There is a high probability that the workflow patterns

exist in real workflow processes, i.e., 60% of the

analyzed workflow processes include organization-

based patterns; 8% include some domain application–

based patterns; and 75% include patterns related to such

business functions;

–

The set of patterns appears to be both necessary and

sufficient to model all the 190 real workflow processes

analyzed. From this, one can conclude that the detected

patterns could be very suitable for defining both

business processes and workflows related to different

application domains. In addition, with few patterns it is

possible to design a large variety of workflows which

we believe can reduce complexity and design effort (see

[4]).

–

A set of rules that not only define specific workflow

patterns but also show how they are combined with

existent control flow patterns (e.g., sequence, XOR-

Split). The rules are described in [4].

We illustrate the approval, unidirectional performative and

the notification pattern as examples.

2.1 Examples of Workflow Patterns

A block activity is suitable to represent each pattern according to

[15]. The block activity concept is particularly suited because it

allows to encapsulate the well-defined semantics and to represent

their atomic characteristics. This means that all activities defined

inside a block activity pattern must be completed before the

superordinated workflow can continue its execution.

Since the patterns representation may require input/output

parameters and the block activity concept does not support

parameters, the transaction perspective of serialization theory was

applied to overcome this limitation [16]. Accordingly, an input

parameter is represented as a database read operation of one-time-

only readable information. Similarly, an output parameter is

represented in the block as a database write operation of one-time-

only writable information.

We describe the three example patterns as an UML Activity

Diagram (using the UML 2.0 notation). The Visual Paradigm for

the UML Community Edition based on UML 2.0 was used as an

editing tool to design the patterns. Figures from 3-5 must be read

according to the legend in Figure 2.

Initial

Node

Atomic

Activity

Decision Node

ForkNode or AND-

Split

Join

Node

Control

Flow

ActivityPartitio

n or Swimlane

Activity

Final

Expansion

region

Figure 2. UML Activity Diagrams

2.1.1 Document Approval Pattern

Context: A doccument must be evaluated by one or more

organizational role.

Problem: How to model a human decision-making in the

workflow process?

Forces:

- The number of times that the decision-making

activity is repeated may vary depending on the

level of centralization of authority (less or more) as

well as the direct supervision of work existent in

the organizational unit(s) where the process is

executed.

- The decision-making activity must be performed

by a human.

- The decision-making activity must have more than

one kind of response (e.g. approval and repproval).

Solution: To include in the workflow, a human activity that

characterizes a point of decision-making on the sub-product in

question (e.g. a document requiring approval). Figure 3 shows the

process construct.

Figure 3. Approval pattern

In Fig. 3 an organizational role reviewer performs a document

review either resulting in an approval or disapproval. The

document review activity is performed multiple times in parallel

or in sequence according to the number of organizational roles

specified or until a disapproval occurs. Generally, the number of

organizational roles is connected to the level of centralization (in

high positions of the organization) with respect to decision-

making.

2.1.2 Unidirectional Performative Pattern

Context: In a workflow, there is a moment that the process must

request the execution of an activity to the system or to an

organizational role involved in the process.

Problem: While modeling the workflow, how to send a request

without waiting for the result of the activity execution?

Forces:

- The receiver’s response is not required.

- The process must keep its execution without

waiting for the activity to be completed.

- The request can be done to the system or a human.

Solution: To include in the workflow model the sequence of

activities (see Fig. 4) representing the unidirectional performative

message. The sequence of activities comprises the generation of a

work item in the receiver worklist. However, the workflow does

not wait for a receiver response to continue execution.

Figure 4. Unidirectional performative message pattern

A sender uses unidirectional performative messages to request the

execution of an activity from a receiver. As shown in Fig. 4, an

activity execution request results in a work item being assigned to

a receiver (i.e., a specific workflow participant responsible for

activity execution). After that, the process may continue execution

without waiting for a response.

2. 1.3 Notification Pattern

Context: During the process execution, some specific events have

more relevancy such that the process must inform some

organizational roles about them.

Problem: While modeling the workflow, how can we keep the

involved roles informed about some process instance events?

Forces:

- The notification must be sent by an electronic way.

- The process does not have to wait for a reading

response to keep its execution.

- The notification must contain the status of a process

activity which the monitoring is desired.

Solution: To include in the workflow the structure concerning the

notifying and do not wait for response. Figure 3 shows how this

structure works. There’s the sending of the notification and its

receiving. These activities must inform the involved roles in the

process about news inherent to the respective workflow, such as,

for example, the approval (or repproval) of a document, or a task

that has achieved its timeout.

Figure 5. Notification pattern

2.2 Core Characteristics of the Workflow

Patterns

We also investigated the frequency of the selected patterns to

specific characteristics of the activities where they were identified.

First, we analyzed the purpose (content) of each activity and

identified the most related pattern (e.g. approval activity 

approval pattern). Afterwards, we annotated and counted the type

of the activity (i.e., automatic or manual). In our last step we

identified the subsequent control flow connected to the activity.

Table 1 summarizes the results of this investigation. It shows, per

example that 97 of the analyzed approval processes (i.e., more

than 85% of the total number of processes) can be defined in

terms of a composition of a bidirectional performative pattern in

a manual activity followed by an Exclusive Choice (XOR-Split)

control flow pattern. Based on such information we defined a set

of rules that connect selected patterns with specific control flows

(see [2]).

Table 1. Specific characteristics of workflow patterns

Kind of

activity Subsequent

control

flow

Frequency (%) of

subsequent control

flow

Unidirectiona

l Manual Sequence Workflows = 142

i.e., 99%

Bi-directional Manual

/automatic XOR-Split Workflows = 123

i.e., 100%

Decision Automatic XOR-Split Workflows = 132

i.e., 92%

Notification Manual

/automatic Sequence Workflows = 102

i.e., 100%

Informative Manual Sequence Workflows = 31

i.e., 100%

Approval manual XOR-Split Workflows = 97

i.e., 85%

3. THE INTELLIGENT WORKFLOW

DESIGNER

This Section describes an approach towards the development of a

suite for both process modeling and normalization based on the

reuse of workflow patterns such as the ones presented in Section

2.1). We intend to use this suite in the extension of some

workflow design tool (e.g., Intalio, Event Process Chain – EPC).

We believe that our patterns can improve the correctness of

workflow design as well as reduce design efforts. In [4] we

present the first inside in this direction.

The core functionalities of the Intelligent Workflow Designer are:

1. Extraction of normalized business processes from legacy

systems: Comprises the extraction of business rules from the

analysis of source code (e.g. COBOL, clipper, access, visual

basic, C++) of legacy systems and subsequent generation of

business processes in high-level notation (such as the

Business Process Modeling Notation - BPMN). The process

is then, validated by matching it with existent workflow

patterns (such as the ones Section 2.1 presents) stored in a

knowledge database. The challenge here is to identify all

embody patterns comprised by the process. As a result the

process is translated into one or more patterns (such as the

ones proposed in [4]). Such procedure must benefit the

translation of the processes to some execution language (e.g.,

BPEL4WS). Furthermore, with the scope of business process

extraction, a model checking is performed, in order to test

the correctness (accuracy) of the process.

2. Support to process design: a user process is received by the

intelligent workflow designer as an input. The process is

then, matched with patterns stored in the knowledge database

in order to identify the partial order of patterns it comprises.

Having this information, the intelligent designer will suggest

the most suitable patterns that are feasible to be used

subsequently by the already designed process. In addition, it

will inform how often each pattern combination was used in

earlier modeling.

3. Construction of a knowledge database of workflow patterns:

The workflow patterns repository (ontology) will store not

only the patterns but also the frequency with each pattern is

most feasible to be combined with other patterns (e.g.,

control flow patterns). Through the mining of new processes

we believe that such frequencies can be improved in order to

increase their precision. Thus, in design time the accuracy,

concerning the frequency associated with each suggestion of

combined patterns be correct may increase. Figure 6

illustrates the suite.

Figure 6. Intelligent Workflow Designer Suite

Core components of the Intelligent Designer Workflow are:

• Legacy Program Flow Extractor (LPFE): component

responsible by the extraction of business process rules

from the source code of legacy systems. Moreover,

generation of corresponding process in high-level

language (such as Business Process Modeling Notation

- BPMN).

• Business Process Model Checking (BPMC): this

component verifies how complete and correct the

extracted process is. First the process is translated to

some formal language (e.g., Pi-calculus). In case it is

correct, the process is matched with the knowledge base

so that the patterns comprised by the process can be

identified.

• Matching Algorithms: algorithms responsible by the

identification (matching) of the workflow patterns

stored in the ontology. The selected patterns are those

comprised by either the user process or the processes

generated by the LPFE and BPMC components.

• Knowledge Base: is the database where the workflow

patterns are stored. It is composed by an ontology which

describes the patterns. It also comprises a query and

update language (mechanism). This mechanism is useful

to identify the most suitable patterns (concerning earlier

use frequency in modeling) to be used subsequently of a

given pattern (based on earlier probabilistic mining

results). In addition, the update mechanism must be

used to change the probabilistic results of each sequence

of patterns based on the process mining results.

• Business Process Mining: External tool to the

Intelligent Workflow Designer which receives a set of

normalized workflow patterns as input. The output is

then, used to update the knowledge database, in special

the use frequency of each pattern.

4. CONCLUSIONS

While workflow patterns were defined for several aspects related

to process execution, the aspect of recurrent business functions is

only partially addressed by existing work. In prior work, we

identified a set of seven workflow patterns that appear necessary

and sufficient to model an extensive set of workflows from

practice. In other work we investigated in how far process

modeling tools can be tailored to provide a direct support for

pattern reuse. Currently, we are also working in the

documentation of the patterns with Pi-calculus. Moreover, we are

analyzing the sequences of workflow patterns in real workflow

processes in order to study the probabilities of the possible

sequences.

In this paper we investigate in how far process modeling tools can

be tailored to provide a direct support for patterns reuse. We

consider that business process can be either designed from scratch

or extracted from legacy systems existent in organizations. Our

contribution is a suite to the analysis and properties verification in

workflow specification (e.g., correctness, completeness, deadlock,

processes equivalence, livelock, model checking). This suite must

be used in the extension or developing of some workflow tool in

order to makes feasible the modeling of business process from

patterns reuse.

The main advantages of this approach can be summarized as

follows: (a) the completeness and necessity of the workflow

patterns for the workflow process design had already been

evidenced in prior work; (b) the suite is tool-independent and can

be adapted for any business process modeling tool; (c) the

business process model checking can be considered as a very

important component which can help in the verification of how

complete and correct is the process that is being designed,

specially through the matching with patterns stored in the

knowledge base.

As future work we not only intend to improve the architecture

presented in this paper but also to implement it through the

development or extension of some workflow design tool. We

strongly believe that the use of the workflow patterns introduced

in this paper must not only reduce design effort but also improve

the correctness of it [3]. Following this trend, we expect to

perform additional experiments concerning process design with

and without the workflow patterns. These experiments are quite

important to demonstrate how process design (from legacy system

or from scratch) can be improved.

ACKNOWLEDGEMENTS

The authors would like to acknowledge the Coordination for the

Improvement of Graduated students (CAPES), the Institute of

Databases and Information Systems

of the University of Ulm (Ulm, Germany) and the Informatics

Institute of Federal University of Rio Grande do Sul (Porto

Alegre, Brazil).

REFERENCES

[1] Thom, L. H.; Iochpe, C.; Amaral, V. L. do; Viero, D. M.

Toward block activity patterns for reuse in workflow design.

In: Workflow Handbook of WfMC, (2006a) pp. 249-260.

[2] Thom, L. H. A Pattern –based Approach for Business

Process Modeling. PPGC: Porto Alegre. Available at:

http://www.biblioteca.ufrgs.br/bibliotecadigital/. (2006b)

(Thesis).

[3] Thom, L. H.; Iochpe, C.; Reichert, M. Workflow Patterns for

Business Process Modeling. In: Workshop on Business

Process Modeling, Development, and Support (BPMDS'07).

Trondheim, Norway. (2007a).

[4] Thom, L. H; Lau, J.M.; Iochpe, C.; Mendling, J. Extending

Business Process Modeling Tools with Workflow Pattern

Reuse In: Proceedings of 9th International Conference on

Enterprise Information Systems (ICEIS 2007), Funchal,

Madeira, Portugal (2007b).

[5] Aalst, W.M.P. van der. Workflow Patterns. In Distributed

and Parallel Databases, 14(3), pages 5-51, (2002).

[6] Russell, N.Workflow Resource Patterns. Brisbane:

Queensland University of Technology. (2004).

[7] Russell, N.; Hofstede, A. H. M Ter; Edmond, D. Workflow

Data Patterns. In: Proceedings of the 24th ER, (2005) pp.

353-368.

[8] Bradshaw, D.; Kennedy, M.; West, C. Oracle BPEL Process

Manager. Developer’s Guide. (2005)

[9] Russell, N.; Aalst, W.M.P. Van Der; Hofstede, A. Ter.

Workflow Exception Patterns. In: Proceedings of 18th

CAiSE, 18, (2006) p.288-302.

[10] Flores, F.; et al. Computer Systems and the Design of

Organizational Interaction. (1998).

[11] Medina-Mora, R. The action workflow approach to

workflow management technology. (1992).

[12] Malone, T. W.; Crownston, K.; Herman, G. A, Organizing

Business Knowledge. (2004).

[13] Muehlen, M. zur. Workflow-based process controlling:

foundations, design, and application of workflow-driven

process information systems. Berlin: Logos Verlag. (2002)

299 p.

[14] Thomas, O.; Scheer, A.-W. Tool Support for the

Collaborative Design of Reference Models - A Business

Engineering Perspective. In: 39th HICSS-39 2006), CD-

ROM / Abstracts Proceedings. (2006).

[15] Workflow Management Coalition. Terminology & Glossary.

Bruxelas, (1999), 65p. Available at: <http://www.wfmc.org>.

Visited on Nov. 2005.

[16] Berbstein, P. A.; Hadzilacos, V.; Goodman, N. Concurrencly

Control and Recovery in Database Systems. Reading:

Addison-Wesley. (1987). 370p.