Information Technology Reference
In-Depth Information
2.1 Foundations
The designed aggregation algorithm inspects an activity environment, i.e., pro-
cess model elements that are related to activities in a process model. Examples
of such elements are data objects accessed by activities and roles supporting ac-
tivity execution, e.g., see model in Fig. 1. The list of such model element types
varies depending on the process modeling language, the tool at hand, modeling
procedures taken into account, and the modeler's style. Each of the model ele-
ment types can be considered as an activity property that has a specific value.
Definition 1 formalizes the activity property concept.
Definition 1 (Activity Property Value and Activity Property Type).
Let
P
be a finite nonempty set of activity property values. Alongside,
T
is a
finite nonempty set of activity property types. Mapping
type
:
P→T
assigns a
type to each value.
The process model in Fig. 1 illustrates Definition 1.
Raw data
,
FA data
,and
Analyst
are examples of activity property values. The process model presents
two activity property types:
Role
and
Data object
. For instance,
type(Raw data)
=Dataobject
,
type(FA data) = Data object
,and
type(Analyst) = Role
.Further,
we define a process model as follows.
Definition 2 (Process Model).
A tuple
m
i
=(
A
i
,G
i
,F
i
,P
i
,props
i
)isa
process model,
where:
-
A
i
is a finite nonempty set of activities;
-
G
i
is a finite set of gateways;
-
N
i
=
A
i
˙
is the set of nodes, where
˙
∪
G
i
∪
denotes a disjoint union of sets;
-
F
i
⊆
N
i
×
N
i
is the flow relation;
-
P
i
⊆P
is a set of activity property values;
-
props
i
:
A
i
→
2
P
i
is a mapping that assigns property values to an activity.
Definition 2 does not make a distinction between different gateway types, since
the future discussion does not make use of them. Mapping
props
i
assigns activity
property values to model activities. Referring to model
m
in the motivating
example of Fig. 1, mapping
props
i
can be illustrated as
props
i
(Collect data)
=
{
Clerk, Raw data
}
. Notice that Definitions 1 and 2 allow to manage the
considered activity property types in a flexible fashion: it is enough to introduce
a new activity property type to set
T
,thevaluesto
P
, and respectively update
mapping
type
. Thereafter, new activity properties can be easily considered within
the activity aggregation. Finally, we postulate the concept of a process model
collection.
Definition 3 (Process Model Collection).
A tuple
c
=(
M, A, P, σ
)isa
process model collection,
where:
-
M
is
a
nonempty
finite
set
of
n
process
models
with
elements
m
i
=(
A
i
,G
i
,F
i
,P
i
,props
i
), where
i
=1
,
2
,...,n
;
-
A
=
˙
∪
i
=1
,
2
,...,n
A
i
is a set of collection activities;
-
P
=
∪
i
=1
,
2
,...,n
P
i
is a set of collection activity property values;
