DERIVING CANONICAL BUSINESS OBJECT OPERATION
NETS FROM PROCESS MODELS
Wang Zhaoxia
1,2,3
, Wang Jianmin
2
, Wen Lijie
2
and Liu Yingbo
2
1
Department of Computer Science and Technology, Tsinghua University, Beijing, China
2
School of Software, Tsinghua University, Beijing, China
3
Department of Logistical Information Engineering, Logistical Engineering University, Chongqing, China
Keywords: Process model, Business object, Object life cycle, Workflow Coloured Petri Net, Operation net.
Abstract: Process model defines the business object operation orders. It is necessary to validate that the business
object operation sequences are consistent with the reference business object life cycle. In this paper we
propose an approach for deriving the canonical business object operation nets from process models which
are modelled with workflow coloured Petri net, i.e. WFCP-net. Our approach consists of 3 steps. First, the
tasks, which access a certain business object, are modelled with task operation nets. Second, the WFCP-net
is rewritten with these task operation nets. Third, the business object operation net is reduced to the
canonical one.
1 INTRODUCTION
Generally, the term business object denotes data
which is handled in a business process (Engels,2001;
Ryndina 2006). Life cycle of a particular business
object describes its whole life process from creation
to death. It provides the object’s state constraint
criteria which define the state pre-condition for an
operation on the object and the transition rule when
the operation is handled. Thus, the manipulating of
the object in a business process should conform to
the state constraint criteria of the object. If the
execution of a task with relevant operations on the
object violates the criteria, it will cause undesired
results, e.g., exceptions, halting, etc. In order to
avoid the above undesired result, a mechanism is
necessary to validate whether the operations on the
object in the process model complies with the state
constraint criteria of the object. To some extent, it is
a problem of model checking (Clarke, E. et.al, 1996)
which includes three parts: system specification,
system model and checking algorithm. The reference
life cycle of the business object is regard as system
specification while the operation diagram of the
object in the process is regard as system model. The
validating algorithm is equal to checking algorithm.
In the rest of this paper, we use term business object
or object to express the data which is handled in a
business process.
A key difficulty of the above problem is how to
derive the object operation diagram from a process.
Currently some researches are to capture the flow of
objects in a process and to describe an object’s
change between tasks. However, these researches
cannot differentiate between the tasks in the process
and the operations on the object. A method to
capture the flow of object is to annotate task with
read/write operation in the process model (Sun et
al., 2006; Lee et al., 2007). However, abstract read
and write operations are not enough to represent the
abundant concrete operations (such as check in,
check out, and release). Other approaches (Engels,
2001; Ryndina 2006; Küster et al., 2007) focus on
the one-to-one relationship between the tasks and the
operations. But in some application domain, e.g.,
product lifecycle management widely used in
manufacturing enterprise, it is not the one-to-one
relationship between the tasks and the operations. In
one word, current process model does not describe
the concrete operation information of objects.
Motivated by the difficulty, we propose an approach
to derive the canonical business object operation net
from a process model by refinement method and
reduction rules.
The remainder of this paper is organized as
follows. Section 2 presents a real-life example which
182
Zhaoxia W., Jianmin W., Lijie W. and Yingbo L. (2009).
DERIVING CANONICAL BUSINESS OBJECT OPERATION NETS FROM PROCESS MODELS.
In Proceedings of the 11th International Conference on Enterprise Information Systems - Information Systems Analysis and Specification, pages
182-187
DOI: 10.5220/0001996801820187
Copyright
c
SciTePress
contains a reference object life cycle and a business
process to illustrate the idea in this paper. In Section
3 relevant concepts are introduced. Section 4
proposes an approach to derive a canonical object
operation net from a process model. Section 5
discusses related work. Finally, conclusions and
future work are shown in Section 6.
2 AN EXAMPLE
In this section, we give an example from TiPLM
system (TiPLM, 2008). TiPLM system is a kind of
PLM (Product Lifecycle Management) system
which has two main functionalities: the product
structure management and the workflow
management.
2.1 A Reference Object Life Cycle
A business object has a life cycle through which it
evolves and consequently transfers among various
states. The standard life cycle of the engineering
data in TiPLM system is shown in Figure 1. States
are indicated by circles (double circle denotes the
accept state). Transition between states is indicated
by directed arc labelled with operation. For instance,
if an object is in the state of being checked out, the
operation check in is enabled. When the operation is
completed, a state transition from the state of being
checked out to the state of being checked in will be
triggered.
Figure 1: A reference object life cyclefor engineering data.
2.2 A Business Process Example
Figure 2 illustrates a design and review process of
engineering drawings in TiPLM system. The key
steps of this process include task designing, task
verifying, task reviewing, task approving and task
releasing. After the engineering drawings are
designed, they will be checked step by step. Once
the checking of them is failed, they will be returned
to task designing for modification.
Figure 2: Design and review process ofengineering
drawings.
For instance, the operation information on the
object of task designing is:
On the one hand, if the object is not existed, the
operation create is executed. When the object is
created, its current state is the state of being checked
out. Then under the condition that the state of the
object is the state of being checked out, designer
generates the 2D drawings or 3D models. Once the
work is completed, the operation check in is
triggered and the object is checked into database.
Then, the state of the object will change to the state
of being checked in. On the other hand, if the design
is returned by the some subsequent task, such as task
verifying, reviewing and approving, the operation
check out is executed and the object’s state changes
from the state of being checked in to the state of
being checked out. Then the designer can verify the
object.
Due to the limit of space, we omit the description
of other tasks. The relationship between tasks in the
process model and operations on the object is shown
in Table 1.
Table 1: Tasks in the process and operations on the object.
task relevant operations on the object
designing create, check in, check out
verifying null
reviewing null
approving null
releasing release
It is apparent that there is no one-to-one
corresponding relation between the tasks and the
operations. Therefore, in order to identify whether
the operation sequence of an object in a business
process is consistent with the reference object life
cycle, a mechanism to characterize the operations on
the object in the process model and to derive the
DERIVING CANONICAL BUSINESS OBJECT OPERATION NETS FROM PROCESS MODELS
183
object operation net is needed. The concrete solution
is elaborated in Section 4.
3 PRELIMINARIES
This section introduces the key concepts relevant
with our work.
3.1 Object Life Cycle
Generally, a non-deterministic finite state machine is
a common means of modelling object life cycle
(Stumpiner, M.& Schrefl, M., 2000). The formal
definition of object life cycle is as follows.
Definition 3.1 (Object Life Cycle). Given an object
O, its object life cycle
,,,
,
where:
is a finite set of operations,
is a finite set of states,
: 2
is the transition function,
mapping each state and operation to a set of
possible successor states,
is the initial state, and
is a set of
accept states.
For example, the reference object life cycle of
business object shown in Figure 1 can be described
as:
1,2,3,4,5,6,7,8,
1, 2,3, 4, 5,6,
=1,
4,6
.
3.2 WFCP-net
As a kind of workflow net, WFCP-net is extended
based on WF-net (Aalst, 1998) and Coloured Petri
Nets (Jensen, 1997). For a detailed introduction to
WFCP-net, the reader is referred to (Liu et al.,
2002).
Definition 3.2 (WFCP-net). A
,,,,,
,, is a WFCP-net if and only if:
(i)
and are subsets of ; which only
has one element is a set of workflow start
places,
which may have one or many
elements is a set of terminal places formal
description:
, :
|
|
1;||
1 and ,·; , · .
(ii)
∉
∉
, is on the
path from
to .
(iii) is category of transition:
:
,
,,,,.
(iv) The initialization function:
,
,
.
To guarantee the process is started and ended
correctly, we add two types of task:
and
.
For instance, the model of the business process
example in Section 2 is shown in Figure 3(a).
4 DERIVING CANONICAL
BUSINESS OBJECT
OPERATION NET
It is noted that we only discuss how to derive the
canonical operation net of a certain object in a
business process. With the similar method, the
corresponding canonical object operation net for any
other objects handled in a business process is also
derived.
4.1 Task Operation Net
In order to characterize the operation information on
a certain object in a process model, the particular
task including operations on the object can be
annotated with object’s operation information. Here,
we propose the task business object operation net
(TBO2-net) to characterize the object operation
information in task. Briefly, we call it task operation
net.
Definition 4.1 (TBO2-net). A TBO2-net is an 8-
tuple
,
,
,
,
,
,
,
, which is a subclass
of WFCP-net. The concepts of all elements are
similar with WFCP-net. In addition, specially
pointing out:
The operation on the object is regard as a
particular task type denoted as
. For
, if the
type of t is
, the corresponding guard function of
represents the precondition of operation enabled.
The expression function on the output arc of
denotes the post state of the object when operation
is performed.
For instance, the TBO2-net of task designing and
releasing of the example in Section 2 are expressed
in Figure 3 (b) respectively.
4.2 Business Object Operation Net
Now, we adopt refinement method (Suzuki &
Murata, 1983; Betous-Almeida & Kanoun, 2004) to
derive the business object operation net (BO2-net).
Definition 4.2 (Refinement). Let WFCP-net
,,,,,,, , TBO2-net
,
,
,
,
,
,
,
,
,
is a task set of
refinable task type in
.
, r
denotes the input
ICEIS 2009 - International Conference on Enterprise Information Systems
184
place of task
,
denotes the output place of task ,
satisfying:
·r
t
r
· ,
|
·t
|
|t·|1 .
corresponding TBO2-net
, is the source place of
, is the sink place of
,
. A new
refinement net
is acquired by using
to
substitute
,
,
,
,
,
,
,
,
.
is
called BO2-net , where:
(i)
.
(ii) Given p
and p
, which are fusion places to
glue WFCP-net and TBO2-net
together.
denotes input fusion place, in
addition,
denotes output fusion place.
Satisfying:
·
·
,
·
.
,
, and
·
· ,
·
·
,
.
(iii)
,
,
,,.
(iv)
.
(v)
,
|
·
,
| ·
,
| ·
,|
·
,
,
,
,
| ·
,|
·
,|
·
,| ·.
(vi)
.
(vii)
.
Figure 3 shows the refinement process. The
BO2-net from the process model of the example is
illustrated in Figure 3(c).
4.3 Canonical Business Object
Operation Net
In this section we introduce reduction rules for
generating the canonical business object operation
net from the business object operation net. The basic
idea is removing the redundant tasks from the
business object operation net and the canonical net
contains pure operation information on the object.
Definition 4.3 (Redundant task). All tasks in a
BO2-net are redundant tasks except the following
task types:
,
and
. Redundant task is
labelled by black rectangle.
The reduction rules are presented in Figure 4.
The first two rules (a) and (b) show that a redundant
task (transition) connecting two places may be
removed by merging the two places, provided that
tokens in the first place can only move to the second
place. The rule (c) shows that a short loop of a
redundant task may be removed. The rule (d) shows
that two parallel redundant tasks can be merged into
one. The rule (e) and (f) show that two redundant
tasks can be combined into one provided that the
place between them has no other links. The last two
rules (g) and (h) are similar to rule (e) and (f) except
the two tasks are both redundant tasks.
Figure 5 shows the reduction process for the
example in Section 2.
The generated net describes the consecutive
operation evolvement of a particular object in the
business process. With existing model checking
tools, e.g. SPIN (Holzmann G.J., 2003), we can
implement the validation strategy. Due to space
limitations, we do not intend to elaborate the
concrete implement process.
5 RELATED WORK
The research area related with our work is
constructing state consistency relations between a
business process and the related data. The concept of
object life cycle comes from object-oriented
technology which concerns about object life cycle
inheritance and requires consistent specialization of
behavior (Kappel & Schrefl, 1991; Schrefl &
Stumptner, 2002). Later, some researches focus on
establishing a link between the business process
models and the object life cycles in business process
management. An approach for generating a process
specification is introduced in ( Küster et al., 2007).
In this literature, a technique is proposed for
generating a business process description from a
given set of objects and their reference object life
cycle. In (Ryndina et al., 2006), an approach is
presented to establish the required consistency, and
then, two consistency notions are defined to verify
the consistency between the generated object life
cycle and the reference object life cycle.
Our work presented in this paper is enlightened
by (Ryndina et al., 2006). The difference of our
work is that the reference object life cycle is isolated
with the process model, therefore, the operations on
an object and the tasks in a process is not one-to-one
corresponding relationship. On the contrary,
Ryndina et al. only considers one-to-one
corresponding relation between the tasks in a
process and the operations on an object. Thus, we
propose a new approach to characterize the
operation information of an object in a process
model. And then, a series of algorithms based on
Petri-nets are presented to derive the canonical
operation net of the business object from the
business processes.
DERIVING CANONICAL BUSINESS OBJECT OPERATION NETS FROM PROCESS MODELS
185
Figure 3: Refinement of a WFCP-net.
Figure 4: Reduction rules for redundant transitions.
(a) label redundant task
t
op
t
op
t
op
t
start
start
t
end
end
op4
t
op
if i=false
if i=true
if j=false
if j=true
if k=true
if k=false
(b) canonical business object operation net
op3
op1
t
op
t
op
op2
t
op
start end
op4
t
op
op1
op3
op2
[state=s1]
[state=s3]
s2
s2
[state=s2]
s3
[state=s3]
s4
[state=s3]
s4
[state=s2]
s3
[state=s3]
s2
[state=s1]
s2
t
start
t
end
Figure 5: Deriving canonical business object operation net (Redundant tasks are shown as black rectangles).
ICEIS 2009 - International Conference on Enterprise Information Systems
186
6 CONCLUSIONS
In this paper, we have proposed an approach to
derive the canonical object operation net. Firstly, we
have described the object operations of a certain task
by task operation net. Then, the business object
operation net has been generated by replacing the
task in process model based on WFCP-net with the
corresponding task operation net. In the end, we
have represented the reduction rules to derive the
canonical object operation net. The approach can
support the object operation information in the
process model and derive the canonical object
operation net from the process model even if the
operations on the object is not one-to-one
corresponding with the tasks in the process model.
As the future work, we intend: (1) to validate the
consistency between the canonical object operation
net from a process model and the reference object
life cycle. (2) to extend the approach to support
operation information description and consistency
analysis of multi-objects processing in a process and
multi-objects in collaboration processes. (3) to
develop a prototype system to implement the
validation strategy.
ACKNOWLEDGEMENTS
We are grateful to Tsinghua InfoTech Co., Ltd for
their business process models accumulated in
TiPLM system. This work is supported by the 973
National Basic Research Program of China (No.
2009CB320700)，the 863 High-Tech Development
Program of China (No. 2007AA040607)，the 863
High-Tech Development Program of China (No.
2008AA042301) and the Program for New Century
Excellent Talents in University.
REFERENCE
Aalst, W. M. P. V. D., 1998. The application of Petri nets
to workflow management. Journal of Circuits, Systems
and Computers, 8(1), 21-66.
Betous-Almeida, C. & Kanoun, K., 2004. Construction
and stepwise refinement of dependability models.
Performance Evaluation. 56, 1-4 (Mar. 2004), 277-306.
Clarke, E., Grumberg, O., and Long, D., 1996. Model
checking. In Proceedings of the NATO Advanced Study
institute on Deductive Program Design (Marktoberdorf,
Germany, July 26 - August 07, 1994). M. Broy, Ed.
Springer-Verlag, New York, Secaucus, NJ, 305-349.
Holzmann G.J., 2003. The SPIN model checker: primer
and reference manual. Addison-Wesley, 2003.
Jensen K., 1997. Colored petri nets : basic concepts,
analysis methods and practical use. Basic Concepts,
Vol. 1, Berlin: Springer Verlag, 1997.
Kappel, G. & Schrefl, M, 1991. Object/behavior diagrams.
In Proceedings of the Seventh International Conference
on Data Engineering (April. 1991). IEEE Computer
Society, Washington, DC, 530-539.
Küster, J.M., Ryndina,K., Gall,H., 2007. Generation of
business process models for object life cycle
compliance,In: Proceeding of BPM2007, Lecture Notes
in Computer Science, vol. 4714, Springer, Berlin, 2007,
165-181.
Liu, D., Wang, J., Chan, S. C., Sun, J., and Zhang, L.,
2002. Modeling workflow processes with colored Petri
nets. Computers in Industry, 49(3)(Dec. 2002), 267-
281.
Lee, G., Eastman, C. M., and Sacks, R., 2007. Eliciting
information for product modeling using process
modeling. Data & Knowledge Engineering, 62(2)(Aug.
2007), 292-307.
Ryndina, K., Küster, J. M., and Gall, H., 2006.
Consistency of business process models and object life
cycles. In: Proceedings of the 1st Workshop on Quality
in Modeling co-located with MoDELS 2006, Technical
report 0627, Technische Universiteit Eindhoven, 2006.
Suzuki, I. & Murata, T., 1983. A method for stepwise
refinement and abstraction of Petri nets. Journal of
Computer and System Sciences, 27(1), 1983, 51-76.
Stumpiner, M., Schrefl, M., 2000. Behavior Consistent
inheritance in UML. In: Conceptual Modeling – ER
2000, LNCSLecture Notes in Computer Science, vol.
1920, Springer, Heidelberg 2000, 527-542.
Schrefl, M. & Stumptner, M., 2002. Behavior-consistent
specialization of object life cycles. ACM Transitions on
Software Engineering and Methodology, 11(1) (Jan.
2002), 92-148.
Sun, S. X., Zhao, J. L., Nunamaker, J. F., and Sheng, O.
R., 2006. Formulating the data-flow perspective for
business process management. Information Systems
Research, 17(4) (Dec. 2006), 374-391.
TiPLM. (2008) www.thit.com.cn/TiPLM/TiPLM.htm.
DERIVING CANONICAL BUSINESS OBJECT OPERATION NETS FROM PROCESS MODELS
187
